Empowering cyber defenders with AI is critical to tilting the cybersecurity balance back in their favor as they battle cybercriminals and keep users safe. To help accelerate adoption of AI for cybersecurity workflows, we partnered with Airbus at DEF CON 33 to host the GenSec Capture the Flag (CTF), dedicated to human-AI collaboration in cybersecurity. Our goal was to create a fun, interactive environment, where participants across various skill levels could explore how AI can accelerate their daily cybersecurity workflows.

The CTF also offered a valuable opportunity for the community to use Sec-Gemini, Google’s experimental Cybersecurity AI, as an optional assistant available in the UI alongside major LLMs. And we received great feedback on Sec-Gemini, with 77% of respondents saying that they had found Sec-Gemini either “very helpful” or “extremely helpful” in assisting them with solving the challenges.  

We want to thank the DEF CON community for the enthusiastic participation and for making this inaugural event a resounding success. The community feedback during the event has been invaluable for understanding how to improve Sec-Gemini, and we are already incorporating some of the lessons learned into the next iteration. 

We are committed to advancing the AI cybersecurity frontier and will continue working with the community to build tools that help protect people online. Stay tuned as we plan to share more research and key learnings from the CTF with the broader community.

To address this gap and help the ecosystem with deploying robust defenses, Google has supported academic research and developed test platforms to analyze DDR5 memory. Our effort has led to the discovery of new attacks and a deeper understanding of Rowhammer on the current DRAM modules, helping to forge the way for further, stronger mitigations.

What is Rowhammer? 

Rowhammer exploits a vulnerability in DRAM. DRAM cells store data as electrical charges, but these electric charges leak over time, causing data corruption. To prevent data loss, the memory controller periodically refreshes the cells. However, if a cell discharges before the refresh cycle, its stored bit may corrupt. Initially considered a reliability issue, it has been leveraged by security researchers to demonstrate privilege escalation attacks. By repeatedly accessing a memory row, an attacker can cause bit flips in neighboring rows. An adversary can exploit Rowhammer via:

1. Reliably cause bit flips by repeatedly accessing adjacent DRAM rows.

2. Coerce other applications or the OS into using these vulnerable memory pages.

3. Target security-sensitive code or data to achieve privilege escalation.

4. Or simply corrupt system’s memory to cause denial of service. 

Previous work has repeatedly demonstrated the possibility of such attacks from software [Revisiting rowhammer, Are we susceptible to rowhammer?, Drammer,  Flip feng shui, Jolt]. As a result, defending against Rowhammer is required for secure isolation in multi-tenant environments like the cloud. 

Rowhammer Mitigations 

The primary approach to mitigate Rowhammer is to detect which memory rows are being aggressively accessed and refreshing nearby rows before a bit flip occurs. TRR is a common example, which uses a number of counters to track accesses to a small number of rows adjacent to a potential victim row. If the access count for these aggressor rows reaches a certain threshold, the system issues a refresh to the victim row. TRR can be incorporated within the DRAM or in the host CPU.

However, this mitigation is not foolproof. For example, the TRRespass attack showed that by simultaneously hammering multiple, non-adjacent rows, TRR can be bypassed. Over the past couple of years, more sophisticated attacks [Half-Double, Blacksmith] have emerged, introducing more efficient attack patterns. 

In response, one of our efforts was to collaborate with JEDEC, external researchers, and experts to define the PRAC as a new mitigation that deterministically detects Rowhammer by tracking all memory rows. 

However, current systems equipped with DDR5 lack support for PRAC or other robust mitigations. As a result, they rely on probabilistic approaches such as ECC and enhanced TRR to reduce the risk. While these measures have mitigated older attacks, their overall effectiveness against new techniques was not fully understood until our recent findings.

Challenges with Rowhammer Assessment 

Mitigating Rowhammer attacks involves making it difficult for an attacker to reliably cause bit flips from software. Therefore, for an effective mitigation, we have to understand how a determined adversary introduces memory accesses that bypass existing mitigations. Three key information components can help with such an analysis:

1. How the improved TRR and in-DRAM ECC work.

2. How memory access patterns from software translate into low-level DDR commands.

3. (Optionally) How any mitigations (e.g., ECC or TRR) in the host processor work.

The first step is particularly challenging and involves reverse-engineering the proprietary in-DRAM TRR mechanism, which varies significantly between different manufacturers and device models. This process requires the ability to issue precise DDR commands to DRAM and analyze its responses, which is difficult on an off-the-shelf system. Therefore, specialized test platforms are essential.

The second and third steps involve analyzing the DDR traffic between the host processor and the DRAM. This can be done using an off-the-shelf interposer, a tool that sits between the processor and DRAM. A crucial part of this analysis is understanding how a live system translates software-level memory accesses into the DDR protocol.

The third step, which involves analyzing host-side mitigations, is sometimes optional. For example, host-side ECC (Error Correcting Code) is enabled by default on servers, while host-side TRR has only been implemented in some CPUs. 

Rowhammer testing platforms

For the first challenge, we partnered with Antmicro to develop two specialized, open-source FPGA-based Rowhammer test platforms. These platforms allow us to conduct in-depth testing on different types of DDR5 modules.

• DDR5 RDIMM Platform: A new DDR5 Tester board to meet the hardware requirements of Registered DIMM (RDIMM) memory, common in server computers.

• SO-DIMM Platform: A version that supports the standard SO-DIMM pinout compatible with off-the-shelf DDR5 SO-DIMM memory sticks, common in workstations and end-user devices.

Antmicro designed and manufactured these open-source platforms and we worked closely with them, and researchers from ETH Zurich, to test the applicability of these platforms for analyzing off-the-shelf memory modules in RDIMM and SO-DIMM forms.

Antmicro DDR5 RDIMM FPGA test platform in action.

Phoenix Attacks on DDR5

In collaboration with researchers from ETH, we applied the new Rowhammer test platforms to evaluate the effectiveness of current in-DRAM DDR5 mitigations. Our findings, detailed in the recently co-authored "Phoenix” research paper, reveal that we successfully developed custom attack patterns capable of bypassing enhanced TRR (Target Row Refresh) defense on DDR5 memory. We were able to create a novel self-correcting refresh synchronization attack technique, which allowed us to perform the first-ever Rowhammer privilege escalation exploit on a standard, production-grade desktop system equipped with DDR5 memory. While this experiment was conducted on an off-the-shelf workstation equipped with recent AMD Zen processors and SK Hynix DDR5 memory, we continue to investigate the applicability of our findings to other hardware configurations.

Lessons learned 

We showed that current mitigations for Rowhammer attacks are not sufficient, and the issue remains a widespread problem across the industry. They do make it more difficult “but not impossible” to carry out attacks, since an attacker needs an in-depth understanding of the specific memory subsystem architecture they wish to target.

Current mitigations based on TRR and ECC rely on probabilistic countermeasures that have insufficient entropy. Once an analyst understands how TRR operates, they can craft specific memory access patterns to bypass it. Furthermore, current ECC schemes were not designed as a security measure and are therefore incapable of reliably detecting errors.

Memory encryption is an alternative countermeasure for Rowhammer. However, our current assessment is that without cryptographic integrity, it offers no valuable defense against Rowhammer. More research is needed to develop viable, practical encryption and integrity solutions.

Path forward

Google has been a leader in JEDEC standardization efforts, for instance with PRAC, a fully approved standard to be supported in upcoming versions of DDR5/LPDDR6. It works by accurately counting the number of times a DRAM wordline is activated and alerts the system if an excessive number of activations is detected. This close coordination between the DRAM and the system gives PRAC a reliable way to address Rowhammer. 

In the meantime, we continue to evaluate and improve other countermeasures to ensure our workloads are resilient against Rowhammer. We collaborate with our academic and industry partners to improve analysis techniques and test platforms, and to share our findings with the broader ecosystem.

Want to learn more?

“Phoenix: Rowhammer Attacks on DDR5 with Self-Correcting Synchronization” will be presented at IEEE Security & Privacy 2026 in San Francisco, CA (MAY 18-21, 2026).

Today we're excited to announce OSS Rebuild, a new project to strengthen trust in open source package ecosystems by reproducing upstream artifacts. As supply chain attacks continue to target widely-used dependencies, OSS Rebuild gives security teams powerful data to avoid compromise without burden on upstream maintainers.

The project comprises:

• Automation to derive declarative build definitions for existing PyPI (Python), npm (JS/TS), and Crates.io (Rust) packages.

• SLSA Provenance for thousands of packages across our supported ecosystems, meeting SLSA Build Level 3 requirements with no publisher intervention.

• Build observability and verification tools that security teams can integrate into their existing vulnerability management workflows.

• Infrastructure definitions to allow organizations to easily run their own instances of OSS Rebuild to rebuild, generate, sign, and distribute provenance.

Challenges

Open source software has become the foundation of our digital world. From critical infrastructure to everyday applications, OSS components now account for 77% of modern applications. With an estimated value exceeding $12 trillion, open source software has never been more integral to the global economy.

Yet this very ubiquity makes open source an attractive target: Recent high-profile supply chain attacks have demonstrated sophisticated methods for compromising widely-used packages. Each incident erodes trust in open ecosystems, creating hesitation among both contributors and consumers.

The security community has responded with initiatives like OpenSSF Scorecard, pypi's Trusted Publishers, and npm's native SLSA support. However, there is no panacea: Each effort targets a certain aspect of the problem, often making tradeoffs like shifting work onto publishers and maintainers.

Our Aim

Our aim with OSS Rebuild is to empower the security community to deeply understand and control their supply chains by making package consumption as transparent as using a source repository. Our rebuild platform unlocks this transparency by utilizing a declarative build process, build instrumentation, and network monitoring capabilities which, within the SLSA Build framework, produces fine-grained, durable, trustworthy security metadata.

Building on the hosted infrastructure model that we pioneered with OSS Fuzz for memory issue detection, OSS Rebuild similarly seeks to use hosted resources to address security challenges in open source, this time aimed at securing the software supply chain.

Our vision extends beyond any single ecosystem: We are committed to bringing supply chain transparency and security to all open source software development. Our initial support for the PyPI (Python), npm (JS/TS), and Crates.io (Rust) package registries—providing rebuild provenance for many of their most popular packages—is just the beginning of our journey.

How OSS Rebuild Works

Through automation and heuristics, we determine a prospective build definition for a target package and rebuild it. We semantically compare the result with the existing upstream artifact, normalizing each one to remove instabilities that cause bit-for-bit comparisons to fail (e.g. archive compression). Once we reproduce the package, we publish the build definition and outcome via SLSA Provenance. This attestation allows consumers to reliably verify a package's origin within the source history, understand and repeat its build process, and customize the build from a known-functional baseline (or maybe even use it to generate more detailed SBOMs).

With OSS Rebuild's existing automation for PyPI, npm, and Crates.io, most packages obtain protection effortlessly without user or maintainer intervention. Where automation isn't currently able to fully reproduce the package, we offer manual build specification so the whole community benefits from individual contributions.

And we are also excited at the potential for AI to help reproduce packages: Build and release processes are often described in natural language documentation which, while difficult to utilize with discrete logic, is increasingly useful to language models. Our initial experiments have demonstrated the approach's viability in automating exploration and testing, with limited human intervention, even in the most complex builds.

Our Capabilities

OSS Rebuild helps detect several classes of supply chain compromise:

• Unsubmitted Source Code - When published packages contain code not present in the public source repository, OSS Rebuild will not attest to the artifact.

• Real world attack: solana/webjs (2024)

• Build Environment Compromise - By creating standardized, minimal build environments with comprehensive monitoring, OSS Rebuild can detect suspicious build activity or avoid exposure to compromised components altogether.

• Real world attack: tj-actions/changed-files (2025)

• Stealthy Backdoors - Even sophisticated backdoors like xz often exhibit anomalous behavioral patterns during builds. OSS Rebuild's dynamic analysis capabilities can detect unusual execution paths or suspicious operations that are otherwise impractical to identify through manual review.

• Real world attack: xz-utils (2024)

For enterprises and security professionals, OSS Rebuild can...

• Enhance metadata without changing registries by enriching data for upstream packages. No need to maintain custom registries or migrate to a new package ecosystem.

• Augment SBOMs by adding detailed build observability information to existing Software Bills of Materials, creating a more complete security picture.

• Accelerate vulnerability response by providing a path to vendor, patch, and re-host upstream packages using our verifiable build definitions.

For publishers and maintainers of open source packages, OSS Rebuild can...

• Strengthen package trust by providing consumers with independent verification of the packages' build integrity, regardless of the sophistication of the original build.

• Retrofit historical packages' integrity with high-quality build attestations, regardless of whether build attestations were present or supported at the time of publication.

• Reduce CI security-sensitivity allowing publishers to focus on core development work. CI platforms tend to have complex authorization and execution models and by performing separate rebuilds, the CI environment no longer needs to be load-bearing for your packages' security.

Check it out!

The easiest (but not only!) way to access OSS Rebuild attestations is to use the provided Go-based command-line interface. It can be compiled and installed easily:

$ go install github.com/google/oss-rebuild/cmd/oss-rebuild@latest

You can fetch OSS Rebuild's SLSA Provenance:

$ oss-rebuild get cratesio syn 2.0.39

..or explore the rebuilt versions of a particular package:

$ oss-rebuild list pypi absl-py

..or even rebuild the package for yourself:

$ oss-rebuild get npm lodash 4.17.20 --output=dockerfile | \

   docker run $(docker buildx build -q -)

Join Us in Helping Secure Open Source

OSS Rebuild is not just about fixing problems; it's about empowering end-users to make open source ecosystems more secure and transparent through collective action. If you're a developer, enterprise, or security researcher interested in OSS security, we invite you to follow along and get involved!

• Check out the code, share your ideas, and voice your feedback at github.com/google/oss-rebuild.

• Explore the data and contribute to improving support for your critical ecosystems and packages.

• Learn more about SLSA Provenance at slsa.dev

With the rapid adoption of generative AI, a new wave of threats is emerging across the industry with the aim of manipulating the AI systems themselves. One such emerging attack vector is indirect prompt injections. Unlike direct prompt injections, where an attacker directly inputs malicious commands into a prompt, indirect prompt injections involve hidden malicious instructions within external data sources. These may include emails, documents, or calendar invites that instruct AI to exfiltrate user data or execute other rogue actions. As more governments, businesses, and individuals adopt generative AI to get more done, this subtle yet potentially potent attack becomes increasingly pertinent across the industry, demanding immediate attention and robust security measures.

At Google, our teams have a longstanding precedent of investing in a defense-in-depth strategy, including robust evaluation, threat analysis, AI security best practices, AI red-teaming, adversarial training, and model hardening for generative AI tools. This approach enables safer adoption of Gemini in Google Workspace and the Gemini app (we refer to both in this blog as “Gemini” for simplicity). Below we describe our prompt injection mitigation product strategy based on extensive research, development, and deployment of improved security mitigations.

A layered security approach

Google has taken a layered security approach introducing security measures designed for each stage of the prompt lifecycle. From Gemini 2.5 model hardening, to purpose-built machine learning (ML) models detecting malicious instructions, to system-level safeguards, we are meaningfully elevating the difficulty, expense, and complexity faced by an attacker. This approach compels adversaries to resort to methods that are either more easily identified or demand greater resources. 

Our model training with adversarial data significantly enhanced our defenses against indirect prompt injection attacks in Gemini 2.5 models (technical details). This inherent model resilience is augmented with additional defenses that we built directly into Gemini, including: 

1. Prompt injection content classifiers

2. Security thought reinforcement

3. Markdown sanitization and suspicious URL redaction

4. User confirmation framework

5. End-user security mitigation notifications

This layered approach to our security strategy strengthens the overall security framework for Gemini – throughout the prompt lifecycle and across diverse attack techniques.

1. Prompt injection content classifiers

Through collaboration with leading AI security researchers via Google's AI Vulnerability Reward Program (VRP), we've curated one of the world’s most advanced catalogs of generative AI vulnerabilities and adversarial data. Utilizing this resource, we built and are in the process of rolling out proprietary machine learning models that can detect malicious prompts and instructions within various formats, such as emails and files, drawing from real-world examples. Consequently, when users query Workspace data with Gemini, the content classifiers filter out harmful data containing malicious instructions, helping to ensure a secure end-to-end user experience by retaining only safe content. For example, if a user receives an email in Gmail that includes malicious instructions, our content classifiers help to detect and disregard malicious instructions, then generate a safe response for the user. This is in addition to built-in defenses in Gmail that automatically block more than 99.9% of spam, phishing attempts, and malware.

A diagram of Gemini’s actions based on the detection of the malicious instructions by content classifiers.

2. Security thought reinforcement

This technique adds targeted security instructions surrounding the prompt content to remind the large language model (LLM) to perform the user-directed task and ignore any adversarial instructions that could be present in the content. With this approach, we steer the LLM to stay focused on the task and ignore harmful or malicious requests added by a threat actor to execute indirect prompt injection attacks.

A diagram of Gemini’s actions based on additional protection provided by the security thought reinforcement technique. 

3. Markdown sanitization and suspicious URL redaction 

Our markdown sanitizer identifies external image URLs and will not render them, making the “EchoLeak” 0-click image rendering exfiltration vulnerability not applicable to Gemini. From there, a key protection against prompt injection and data exfiltration attacks occurs at the URL level. With external data containing dynamic URLs, users may encounter unknown risks as these URLs may be designed for indirect prompt injections and data exfiltration attacks. Malicious instructions executed on a user's behalf may also generate harmful URLs. With Gemini, our defense system includes suspicious URL detection based on Google Safe Browsing to differentiate between safe and unsafe links, providing a secure experience by helping to prevent URL-based attacks. For example, if a document contains malicious URLs and a user is summarizing the content with Gemini, the suspicious URLs will be redacted in Gemini’s response. 

Gemini in Gmail provides a summary of an email thread. In the summary, there is an unsafe URL. That URL is redacted in the response and is replaced with the text “suspicious link removed”. 

4. User confirmation framework

Gemini also features a contextual user confirmation system. This framework enables Gemini to require user confirmation for certain actions, also known as “Human-In-The-Loop” (HITL), using these responses to bolster security and streamline the user experience. For example, potentially risky operations like deleting a calendar event may trigger an explicit user confirmation request, thereby helping to prevent undetected or immediate execution of the operation.

The Gemini app with instructions to delete all events on Saturday. Gemini responds with the events found on Google Calendar and asks the user to confirm this action.

5. End-user security mitigation notifications

A key aspect to keeping our users safe is sharing details on attacks that we’ve stopped so users can watch out for similar attacks in the future. To that end, when security issues are mitigated with our built-in defenses, end users are provided with contextual information allowing them to learn more via dedicated help center articles. For example, if Gemini summarizes a file containing malicious instructions and one of Google’s prompt injection defenses mitigates the situation, a security notification with a “Learn more” link will be displayed for the user. Users are encouraged to become more familiar with our prompt injection defenses by reading the Help Center article. 

Gemini in Docs with instructions to provide a summary of a file. Suspicious content was detected and a response was not provided. There is a yellow security notification banner for the user and a statement that Gemini’s response has been removed, with a “Learn more” link to a relevant Help Center article.

Moving forward

Our comprehensive prompt injection security strategy strengthens the overall security framework for Gemini. Beyond the techniques described above, it also involves rigorous testing through manual and automated red teams, generative AI security BugSWAT events, strong security standards like our Secure AI Framework (SAIF), and partnerships with both external researchers via the Google AI Vulnerability Reward Program (VRP) and industry peers via the Coalition for Secure AI (CoSAI). Our commitment to trust includes collaboration with the security community to responsibly disclose AI security vulnerabilities, share our latest threat intelligence on ways we see bad actors trying to leverage AI, and offering insights into our work to build stronger prompt injection defenses. 

Working closely with industry partners is crucial to building stronger protections for all of our users. To that end, we’re fortunate to have strong collaborative partnerships with numerous researchers, such as Ben Nassi (Confidentiality), Stav Cohen (Technion), and Or Yair (SafeBreach), as well as other AI Security researchers participating in our BugSWAT events and AI VRP program. We appreciate the work of these researchers and others in the community to help us red team and refine our defenses.

We continue working to make upcoming Gemini models inherently more resilient and add additional prompt injection defenses directly into Gemini later this year. To learn more about Google’s progress and research on generative AI threat actors, attack techniques, and vulnerabilities, take a look at the following resources:

• Beyond Speculation: Data-Driven Insights into AI and Cybersecurity (RSAC 2025 conference keynote) from Google’s Threat Intelligence Group (GTIG)

• Adversarial Misuse of Generative AI (blog post) from Google’s Threat Intelligence Group (GTIG)

• Google's Approach for Secure AI Agents (white paper) from Google’s Secure AI Framework (SAIF) team

• Advancing Gemini's security safeguards (blog post) from Google’s DeepMind team

• Lessons from Defending Gemini Against Indirect Prompt Injections (white paper) from Google’s DeepMind team

Google Quantum AI's mission is to build best in class quantum computing for otherwise unsolvable problems. For decades the quantum and security communities have also known that large-scale quantum computers will at some point in the future likely be able to break many of today’s secure public key cryptography algorithms, such as Rivest–Shamir–Adleman (RSA). Google has long worked with the U.S. National Institute of Standards and Technology (NIST) and others in government, industry, and academia to develop and transition to post-quantum cryptography (PQC), which is expected to be resistant to quantum computing attacks. As quantum computing technology continues to advance, ongoing multi-stakeholder collaboration and action on PQC is critical.

In order to plan for the transition from today’s cryptosystems to an era of PQC, it's important the size and performance of a future quantum computer that could likely break current cryptography algorithms is carefully characterized. Yesterday, we published a preprint demonstrating that 2048-bit RSA encryption could theoretically be broken by a quantum computer with 1 million noisy qubits running for one week. This is a 20-fold decrease in the number of qubits from our previous estimate, published in 2019. Notably, quantum computers with relevant error rates currently have on the order of only 100 to 1000 qubits, and the National Institute of Standards and Technology (NIST) recently released standard PQC algorithms that are expected to be resistant to future large-scale quantum computers. However, this new result does underscore the importance of migrating to these standards in line with NIST recommended timelines. 

Estimated resources for factoring have been steadily decreasing

Quantum computers break RSA by factoring numbers, using Shor’s algorithm. Since Peter Shor published this algorithm in 1994, the estimated number of qubits needed to run it has steadily decreased. For example, in 2012, it was estimated that a 2048-bit RSA key could be broken by a quantum computer with a billion physical qubits. In 2019, using the same physical assumptions – which consider qubits with a slightly lower error rate than Google Quantum AI’s current quantum computers – the estimate was lowered to 20 million physical qubits.

Historical estimates of the number of physical qubits needed to factor 2048-bit RSA integers.

This result represents a 20-fold decrease compared to our estimate from 2019

The reduction in physical qubit count comes from two sources: better algorithms and better error correction – whereby qubits used by the algorithm ("logical qubits") are redundantly encoded across many physical qubits, so that errors can be detected and corrected.

On the algorithmic side, the key change is to compute an approximate modular exponentiation rather than an exact one. An algorithm for doing this, while using only small work registers, was discovered in 2024 by Chevignard and Fouque and Schrottenloher. Their algorithm used 1000x more operations than prior work, but we found ways to reduce that overhead down to 2x.

On the error correction side, the key change is tripling the storage density of idle logical qubits by adding a second layer of error correction. Normally more error correction layers means more overhead, but a good combination was discovered by the Google Quantum AI team in 2023. Another notable error correction improvement is using "magic state cultivation", proposed by the Google Quantum AI team in 2024, to reduce the workspace required for certain basic quantum operations. These error correction improvements aren't specific to factoring and also reduce the required resources for other quantum computations like in chemistry and materials simulation.

Security implications

NIST recently concluded a PQC competition that resulted in the first set of PQC standards. These algorithms can already be deployed to defend against quantum computers well before a working cryptographically relevant quantum computer is built. 

To assess the security implications of quantum computers, however, it’s instructive to additionally take a closer look at the affected algorithms (see here for a detailed look): RSA and Elliptic Curve Diffie-Hellman. As asymmetric algorithms, they are used for encryption in transit, including encryption for messaging services, as well as digital signatures (widely used to prove the authenticity of documents or software, e.g. the identity of websites). For asymmetric encryption, in particular encryption in transit, the motivation to migrate to PQC is made more urgent due to the fact that an adversary can collect ciphertexts, and later decrypt them once a quantum computer is available, known as a “store now, decrypt later” attack. Google has therefore been encrypting traffic both in Chrome and internally, switching to the standardized version of ML-KEM once it became available. Notably not affected is symmetric cryptography, which is primarily deployed in encryption at rest, and to enable some stateless services.

For signatures, things are more complex. Some signature use cases are similarly urgent, e.g., when public keys are fixed in hardware. In general, the landscape for signatures is mostly remarkable due to the higher complexity of the transition, since signature keys are used in many different places, and since these keys tend to be longer lived than the usually ephemeral encryption keys. Signature keys are therefore harder to replace and much more attractive targets to attack, especially when compute time on a quantum computer is a limited resource. This complexity likewise motivates moving earlier rather than later. To enable this, we have added PQC signature schemes in public preview in Cloud KMS. 

The initial public draft of the NIST internal report on the transition to post-quantum cryptography standards states that vulnerable systems should be deprecated after 2030 and disallowed after 2035. Our work highlights the importance of adhering to this recommended timeline.

More from Google on PQC: https://cloud.google.com/security/resources/post-quantum-cryptography?e=48754805 

Today, we’re announcing Sec-Gemini v1, a new experimental AI model focused on advancing cybersecurity AI frontiers. 

As outlined a year ago, defenders face the daunting task of securing against all cyber threats, while attackers need to successfully find and exploit only a single vulnerability. This fundamental asymmetry has made securing systems extremely difficult, time consuming and error prone. AI-powered cybersecurity workflows have the potential to help shift the balance back to the defenders by force multiplying cybersecurity professionals like never before.

Effectively powering SecOps workflows requires state-of-the-art reasoning capabilities and extensive current cybersecurity knowledge. Sec-Gemini v1 achieves this by combining Gemini’s advanced capabilities with near real-time cybersecurity knowledge and tooling. This combination allows it to achieve superior performance on key cybersecurity workflows, including incident root cause analysis, threat analysis, and vulnerability impact understanding.

We firmly believe that successfully pushing AI cybersecurity frontiers to decisively tilt the balance in favor of the defenders requires a strong collaboration across the cybersecurity community. This is why we are making Sec-Gemini v1 freely available to select organizations, institutions, professionals, and NGOs for research purposes.

Sec-Gemini v1 outperforms other models on key cybersecurity benchmarks as a result of its advanced integration of Google Threat Intelligence (GTI), OSV, and other key data sources. Sec-Gemini v1 outperforms other models on CTI-MCQ, a leading threat intelligence benchmark, by at least 11% (See Figure 1). It also outperforms other models by at least 10.5% on the CTI-Root Cause Mapping benchmark (See Figure 2):

Figure 1: Sec-Gemini v1 outperforms other models on the CTI-MCQ Cybersecurity Threat Intelligence benchmark.

Figure 2: Sec-Gemini v1 has outperformed other models in a Cybersecurity Threat Intelligence-Root Cause Mapping (CTI-RCM) benchmark that evaluates an LLM's ability to understand the nuances of vulnerability descriptions, identify vulnerabilities underlying root causes, and accurately classify them according to the CWE taxonomy.

Below is an example of the comprehensiveness of Sec-Gemini v1’s answers in response to key cybersecurity questions. First, Sec-Gemini v1 is able to determine that Salt Typhoon is a threat actor (not all models do) and provides a comprehensive description of that threat actor, thanks to its deep integration with Mandiant Threat intelligence data.

Next, in response to a question about the vulnerabilities in the Salt Typhoon description, Sec-Gemini v1 outputs not only vulnerability details (thanks to its integration with OSV data, the open-source vulnerabilities database operated by Google), but also contextualizes the vulnerabilities with respect to threat actors (using Mandiant data). With Sec-Gemini v1, analysts can understand the risk and threat profile associated with specific vulnerabilities faster.

In partnership with NVIDIA and HiddenLayer, as part of the Open Source Security Foundation, we are now launching the first stable version of our model signing library. Using digital signatures like those from Sigstore, we allow users to verify that the model used by the application is exactly the model that was created by the developers. In this blog post we will illustrate why this release is important from Google’s point of view.

With the advent of LLMs, the ML field has entered an era of rapid evolution. We have seen remarkable progress leading to weekly launches of various applications which incorporate ML models to perform tasks ranging from customer support, software development, and even performing security critical tasks.

However, this has also opened the door to a new wave of security threats. Model and data poisoning, prompt injection, prompt leaking and prompt evasion are just a few of the risks that have recently been in the news. Garnering less attention are the risks around the ML supply chain process: since models are an uninspectable collection of weights (sometimes also with arbitrary code), an attacker can tamper with them and achieve significant impact to those using the models. Users, developers, and practitioners need to examine an important question during their risk assessment process: “can I trust this model?”

Since its launch, Google’s Secure AI Framework (SAIF) has created guidance and technical solutions for creating AI applications that users can trust. A first step in achieving trust in the model is to permit users to verify its integrity and provenance, to prevent tampering across all processes from training to usage, via cryptographic signing. 

The ML supply chain

To understand the need for the model signing project, let’s look at the way ML powered applications are developed, with an eye to where malicious tampering can occur.

Applications that use advanced AI models are typically developed in at least three different stages. First, a large foundation model is trained on large datasets. Next, a separate ML team finetunes the model to make it achieve good performance on application specific tasks. Finally,  this fine-tuned model is embedded into an application.

The three steps involved in building an application that uses large language models.

These three stages are usually handled by different teams, and potentially even different companies, since each stage requires specialized expertise. To make models available from one stage to the next, practitioners leverage model hubs, which are repositories for storing models. Kaggle and HuggingFace are popular open source options, although internal model hubs could also be used.

This separation into stages creates multiple opportunities where a malicious user (or external threat actor who has compromised the internal infrastructure) could tamper with the model. This could range from just a slight alteration of the model weights that control model behavior, to injecting architectural backdoors — completely new model behaviors and capabilities that could be triggered only on specific inputs. It is also possible to exploit the serialization format and inject arbitrary code execution in the model as saved on disk — our whitepaper on AI supply chain integrity goes into more details on how popular model serialization libraries could be exploited. The following diagram summarizes the risks across the ML supply chain for developing a single model, as discussed in the whitepaper.

The supply chain diagram for building a single model, illustrating some supply chain risks (oval labels) and where model signing can defend against them (check marks)

The diagram shows several places where the model could be compromised. Most of these could be prevented by signing the model during training and verifying integrity before any usage, in every step: the signature would have to be verified when the model gets uploaded to a model hub, when the model gets selected to be deployed into an application (embedded or via remote APIs) and when the model is used as an intermediary during another training run. Assuming the training infrastructure is trustworthy and not compromised, this approach guarantees that each model user can trust the model.

Sigstore for ML models

Signing models is inspired by code signing, a critical step in traditional software development. A signed binary artifact helps users identify its producer and prevents tampering after publication. The average developer, however, would not want to manage keys and rotate them on compromise.

These challenges are addressed by using Sigstore, a collection of tools and services that make code signing secure and easy. By binding an OpenID Connect token to a workload or developer identity, Sigstore alleviates the need to manage or rotate long-lived secrets. Furthermore, signing is made transparent so signatures over malicious artifacts could be audited in a public transparency log, by anyone. This ensures that split-view attacks are not possible, so any user would get the exact same model. These features are why we recommend Sigstore’s signing mechanism as the default approach for signing ML models.

Today the OSS community is releasing the v1.0 stable version of our model signing library as a Python package supporting Sigstore and traditional signing methods. This model signing library is specialized to handle the sheer scale of ML models (which are usually much larger than traditional software components), and handles signing models represented as a directory tree. The package provides CLI utilities so that users can sign and verify model signatures for individual models. The package can also be used as a library which we plan to incorporate directly into model hub upload flows as well as into ML frameworks.

Future goals

We can view model signing as establishing the foundation of trust in the ML ecosystem. We envision extending this approach to also include datasets and other ML-related artifacts. Then, we plan to build on top of signatures, towards fully tamper-proof metadata records, that can be read by both humans and machines. This has the potential to automate a significant fraction of the work needed to perform incident response in case of a compromise in the ML world. In an ideal world, an ML developer would not need to perform any code changes to the training code, while the framework itself would handle model signing and verification in a transparent manner.

We’re excited to announce that starting today, Titan Security Keys are available for purchase in more than 10 new countries:

• Ireland

• Portugal

• The Netherlands

• Denmark

• Norway

• Sweden

• Finland

• Australia

• New Zealand

• Singapore

• Puerto Rico

This expansion means Titan Security Keys are now available in 22 markets, including previously announced countries like Austria, Belgium, Canada, France, Germany, Italy, Japan, Spain, Switzerland, the UK, and the US.

What is a Titan Security Key?

A Titan Security Key is a small, physical device that you can use to verify your identity when you sign in to your Google Account. It’s like a second password that’s much harder for cybercriminals to steal.

Titan Security Keys allow you to store your passkeys on a strong, purpose-built device that can help protect you against phishing and other online attacks. They’re easy to use and work with a wide range of devices and services as they’re compatible with the FIDO2 standard.

How do I use a Titan Security Key?

To use a Titan Security Key, you simply plug it into your computer’s USB port or tap it to your device using NFC. When you’re asked to verify your identity, you’ll just need to tap the button on the key.

Where can I buy a Titan Security Key?

You can buy Titan Security Keys on the Google Store.

We’re committed to making our products available to as many people as possible and we hope this expansion will help more people stay safe online.

In December 2022, we released the open source OSV-Scanner tool, and earlier this year, we open sourced OSV-SCALIBR. OSV-Scanner and OSV-SCALIBR, together with OSV.dev are components of an open platform for managing vulnerability metadata and enabling simple and accurate matching and remediation of known vulnerabilities. Our goal is to simplify and streamline vulnerability management for developers and security teams alike.

Today, we're thrilled to announce the launch of OSV-Scanner V2.0.0, following the announcement of the beta version. This V2 release builds upon the foundation we laid with OSV-SCALIBR and adds significant new capabilities to OSV-Scanner, making it a comprehensive vulnerability scanner and remediation tool with broad support for formats and ecosystems. 

What’s new

Enhanced Dependency Extraction with OSV-SCALIBR

This release represents the first major integration of OSV-SCALIBR features into OSV-Scanner, which is now the official command-line code and container scanning tool for the OSV-SCALIBR library. This integration also expanded our support for the kinds of dependencies we can extract from projects and containers:

Source manifests and lockfiles:

• .NET: deps.json

• Python: uv.lock

• JavaScript: bun.lock

• Haskell: cabal.project.freeze, stack.yaml.lock

Artifacts:

• Node modules

• Python wheels

• Java uber jars

• Go binaries

Layer and base image-aware container scanning

Previously, OSV-Scanner focused on scanning of source repositories and language package manifests and lockfiles. OSV-Scanner V2 adds support for comprehensive, layer-aware scanning for Debian, Ubuntu, and Alpine container images. OSV-Scanner can now analyze container images to provide:

• Layers where a package was first introduced

• Layer history and commands

• Base images the image is based on (leveraging a new experimental API provided by deps.dev).

• OS/Distro the container is running on

• Filtering of vulnerabilities that are unlikely to impact your container image

This layer analysis currently supports the following OSes and languages:

Distro Support:

• Alpine OS

• Debian

• Ubuntu

Language Artifacts Support:

• Go

• Java

• Node

• Python

Interactive HTML output

Presenting vulnerability scan information in a clear and actionable way is difficult, particularly in the context of container scanning. To address this, we built a new interactive local HTML output format. This provides more interactivity and information compared to terminal only outputs, including:

• Severity breakdown

• Package and ID filtering

• Vulnerability importance filtering

• Full vulnerability advisory entries

And additionally for container image scanning:

• Layer filtering

• Image layer information

• Base image identification

Illustration of HTML output for container image scanning

Guided remediation for Maven pom.xml

Last year we released a feature called guided remediation for npm, which streamlines vulnerability management by intelligently suggesting prioritized, targeted upgrades and offering flexible strategies. This ultimately maximizes security improvements while minimizing disruption. We have now expanded this feature to Java through support for Maven pom.xml.

With guided remediation support for Maven, you can remediate vulnerabilities in both direct and transitive dependencies through direct version updates or overriding versions through dependency management.

We’ve introduced a few new things for our Maven support:

• A new remediation strategy override.

• Support for reading and writing pom.xml files, including writing changes to local parent pom files. We leverage OSV-Scalibr for Maven transitive dependency extraction.

• A private registry can be specified to fetch Maven metadata.

• A new experimental subcommend to update all your dependencies in pom.xml to the latest version.

We also introduced machine readable output for guided remediation that makes it easier to integrate guided remediation into your workflow.

What’s next?

We have exciting plans for the remainder of the year, including:

• Continued OSV-SCALIBR Convergence: We will continue to converge OSV-Scanner and OSV-SCALIBR to bring OSV-SCALIBR’s functionality to OSV-Scanner’s CLI interface.

• Expanded Ecosystem Support: We'll expand the number of ecosystems we support across all the features currently in OSV-Scanner, including more languages for guided remediation, OS advisories for container scanning, and more general lockfile support for source code scanning.

• Full Filesystem Accountability for Containers: Another goal of osv-scanner is to give you the ability to know and account for every single file on your container image, including sideloaded binaries downloaded from the internet.

• Reachability Analysis: We're working on integrating reachability analysis to provide deeper insights into the potential impact of vulnerabilities.

• VEX Support: We're planning to add support for Vulnerability Exchange (VEX) to facilitate better communication and collaboration around vulnerability information.

Try OSV-Scanner V2

You can try V2.0.0 and contribute to its ongoing development by checking out OSV-Scanner or the OSV-SCALIBR repository. We welcome your feedback and contributions as we continue to improve the platform and make vulnerability management easier for everyone.

In 2024, our Vulnerability Reward Program confirmed the ongoing value of engaging with the security research community to make Google and its products safer. This was evident as we awarded just shy of $12 million to over 600 researchers based in countries around the globe across all of our programs.

Vulnerability Reward Program 2024 in Numbers

You can learn about who’s reporting to the Vulnerability Reward Program via our Leaderboard – and find out more about our youngest security researchers who’ve recently joined the ranks of Google bug hunters.

VRP Highlights in 2024

In 2024 we made a series of changes and improvements coming to our vulnerability reward programs and related initiatives:

• The Google VRP revamped its reward structure, bumping rewards up to a maximum of $151,515, the Mobile VRP is now offering up to $300,000 for critical vulnerabilities in top-tier apps, Cloud VRP has a top-tier award of up $151,515, and Chrome awards now peak at $250,000 (see the below section on Chrome for details).

• We rolled out InternetCTF – to get rewarded, discover novel code execution vulnerabilities in open source and provide Tsunami plugin patches for them.

• The Abuse VRP saw a 40% YoY increase in payouts – we received over 250 valid bugs targeting abuse and misuse issues in Google products, resulting in over $290,000 in rewards.

• To improve the payment process for rewards going to bug hunters, we introduced Bugcrowd as an additional payment option on bughunters.google.com alongside the existing standard Google payment option. 

• We hosted two editions of bugSWAT for training, skill sharing, and, of course, some live hacking – in August, we had 16 bug hunters in attendance in Las Vegas, and in October, as part of our annual security conference ESCAL8 in Malaga, Spain, we welcomed 40 of our top researchers. Between these two events, our bug hunters were rewarded $370,000 (and plenty of swag).

• We doubled down on our commitment to support the next generation of security engineers by hosting four init.g workshops (Las Vegas, São Paulo, Paris, and Malaga). Follow the Google VRP channel on X to stay tuned on future events.

More detailed updates on selected programs are shared in the following sections.

Android and Google Devices

In 2024, the Android and Google Devices Security Reward Program and the Google Mobile Vulnerability Reward Program, both part of the broader Google Bug Hunters program, continued their mission to fortify the Android ecosystem, achieving new heights in both impact and severity. We awarded over $3.3 million in rewards to researchers who demonstrated exceptional skill in uncovering critical vulnerabilities within Android and Google mobile applications. 

The above numbers mark a significant change compared to previous years. Although we saw an 8% decrease in the total number of submissions, there was a 2% increase in the number of critical and high vulnerabilities. In other words, fewer researchers are submitting fewer, but more impactful bugs, and are citing the improved security posture of the Android operating system as the central challenge. This showcases the program's sustained success in hardening Android.

This year, we had a heightened focus on Android Automotive OS and WearOS, bringing actual automotive devices to multiple live hacking events and conferences. At ESCAL8, we hosted a live-hacking challenge focused on Pixel devices, resulting in over $75,000 in rewards in one weekend, and the discovery of several memory safety vulnerabilities. To facilitate learning, we launched a new Android hacking course in collaboration with external security researchers, focused on mobile app security, designed for newcomers and veterans alike. Stay tuned for more.

We extend our deepest gratitude to the dedicated researchers who make the Android ecosystem safer. We're proud to work with you! Special thanks to Zinuo Han (@ele7enxxh) for their expertise in Bluetooth security, blunt (@blunt_qian) for holding the record for the most valid reports submitted to the Google Play Security Reward Program, and WANG,YONG (@ThomasKing2014) for groundbreaking research on rooting Android devices with kernel MTE enabled. We also appreciate all researchers who participated in last year's bugSWAT event in Málaga. Your contributions are invaluable! 

Chrome

Chrome did some remodeling in 2024 as we updated our reward amounts and structure to incentivize deeper research. For example, we increased our maximum reward for a single issue to $250,000 for demonstrating RCE in the browser or other non-sandboxed process, and more if done directly without requiring a renderer compromise. 

In 2024, UAF mitigation MiraclePtr was fully launched across all platforms, and a year after the initial launch, MiraclePtr-protected bugs are no longer being considered exploitable security bugs. In tandem, we increased the MiraclePtr Bypass Reward to $250,128. Between April and November, we also launched the first and second iterations of the V8 Sandbox Bypass Rewards as part of the progression towards the V8 sandbox, eventually becoming a security boundary in Chrome. 

We received 337 reports of unique, valid security bugs in Chrome during 2024, and awarded 137 Chrome VRP researchers $3.4 million in total. The highest single reward of 2024 was $100,115 and was awarded to Mickey for their report of a MiraclePtr Bypass after MiraclePtr was initially enabled across most platforms in Chrome M115 in 2023. We rounded out the year by announcing the top 20 Chrome VRP researchers for 2024, all of whom were gifted new Chrome VRP swag, featuring our new Chrome VRP mascot, Bug.

Cloud VRP

The Cloud VRP launched in October as a Cloud-focused vulnerability reward program dedicated to Google Cloud products and services. As part of the launch, we also updated our product tiering and improved our reward structure to better align our reports with their impact on Google Cloud. This resulted in over 150 Google Cloud products coming under the top two reward tiers, enabling better rewards for our Cloud researchers and a more secure cloud.

Since its launch, Google Cloud VRP triaged over 400 reports and filed over 200 unique security vulnerabilities for Google Cloud products and services leading to over $500,000 in researcher rewards. 

Our highlight last year was launching at the bugSWAT event in Málaga where we got to meet many of our amazing researchers who make our program so successful! The overwhelming positive feedback from the researcher community continues to propel us to mature Google Cloud VRP further this year. Stay tuned for some exciting announcements!

Generative AI

We’re celebrating an exciting first year of AI bug bounties.  We received over 150 bug reports – over $55,000 in rewards so far – with one-in-six leading to key improvements. 

We also ran a bugSWAT live-hacking event targeting LLM products and received 35 reports, totaling more than $87,000 – including issues like “Hacking Google Bard - From Prompt Injection to Data Exfiltration” and “We Hacked Google A.I. for $50,000”.

Keep an eye on Gen AI in 2025 as we focus on expanding scope and sharing additional ways for our researcher community to contribute. 

Looking Forward to 2025

In 2025, we will be celebrating 15 years of VRP at Google, during which we have remained fully committed to fostering collaboration, innovation, and transparency with the security community, and will continue to do so in the future. Our goal remains to stay ahead of emerging threats, adapt to evolving technologies, and continue to strengthen the security posture of Google’s products and services. 

We want to send a huge thank you to our bug hunter community for helping us make Google products and platforms more safe and secure for our users around the world – and invite researchers not yet engaged with the Vulnerability Reward Program to join us in our mission to keep Google safe! 

Thank you to Dirk Göhmann, Amy Ressler, Eduardo Vela, Jan Keller, Krzysztof Kotowicz, Martin Straka, Michael Cote, Mike Antares, Sri Tulasiram, and Tony Mendez.

Tip: Want to be informed of new developments and events around our Vulnerability Reward Program? Follow the Google VRP channel on X to stay in the loop and be sure to check out the Security Engineering blog, which covers topics ranging from VRP updates to security practices and vulnerability descriptions (30 posts in 2024)!

For decades, memory safety vulnerabilities have been at the center of various security incidents across the industry, eroding trust in technology and costing billions. Traditional approaches, like code auditing, fuzzing, and exploit mitigations – while helpful – haven't been enough to stem the tide, while incurring an increasingly high cost.

In this blog post, we are calling for a fundamental shift: a collective commitment to finally eliminate this class of vulnerabilities, anchored on secure-by-design practices – not just for ourselves but for the generations that follow.

The shift we are calling for is reinforced by a recent ACM article calling to standardize memory safety we took part in releasing with academic and industry partners. It's a recognition that the lack of memory safety is no longer a niche technical problem but a societal one, impacting everything from national security to personal privacy.

The standardization opportunity

Over the past decade, a confluence of secure-by-design advancements has matured to the point of practical, widespread deployment. This includes memory-safe languages, now including high-performance ones such as Rust, as well as safer language subsets like Safe Buffers for C++. 

These tools are already proving effective. In Android for example, the increasing adoption of memory-safe languages like Kotlin and Rust in new code has driven a significant reduction in vulnerabilities.

Looking forward, we're also seeing exciting and promising developments in hardware. Technologies like ARM's Memory Tagging Extension (MTE) and the Capability Hardware Enhanced RISC Instructions (CHERI) architecture offer a complementary defense, particularly for existing code.

While these advancements are encouraging, achieving comprehensive memory safety across the entire software industry requires more than just individual technological progress:  we need to create the right environment and accountability for their widespread adoption. Standardization is key to this. 

To facilitate standardization, we suggest establishing a common framework for specifying and objectively assessing memory safety assurances; doing so will lay the foundation for creating a market in which vendors are incentivized to invest in memory safety. Customers will be empowered to recognize, demand, and reward safety. This framework will provide governments and businesses with the clarity to specify memory safety requirements, driving the procurement of more secure systems. 

The framework we are proposing would complement existing efforts by defining specific, measurable criteria for achieving different levels of memory safety assurance across the industry. In this way, policymakers will gain the technical foundation to craft effective policy initiatives and incentives promoting memory safety.

A blueprint for a memory-safe future

We know there's more than one way of solving this problem, and we are ourselves investing in several. Importantly, our vision for achieving memory safety through standardization focuses on defining the desired outcomes rather than locking ourselves into specific technologies.

To translate this vision into an effective standard, we need a framework that will:

Foster innovation and support diverse approaches: The standard should focus on the security properties we want to achieve (e.g., freedom from spatial and temporal safety violations) rather than mandating specific implementation details. The framework should therefore be technology-neutral, allowing vendors to choose the best approach for their products and requirements. This encourages innovation and allows software and hardware manufacturers to adopt the best solutions as they emerge.

Tailor memory safety requirements based on need: The framework should establish different levels of safety assurance, akin to SLSA levels, recognizing that different applications have different security needs and cost constraints. Similarly, we likely need distinct guidance for developing new systems and improving existing codebases. For instance, we probably do not need every single piece of code to be formally proven. This allows for tailored security, ensuring appropriate levels of memory safety for various contexts. 

Enable objective assessment: The framework should define clear criteria and potentially metrics for assessing memory safety and compliance with a given level of assurance. The goal would be to objectively compare the memory safety assurance of different software components or systems, much like we assess energy efficiency today. This will move us beyond subjective claims and towards objective and comparable security properties across products.

Be practical and actionable: Alongside the technology-neutral framework, we need best practices for existing technologies. The framework should provide guidance on how to effectively leverage specific technologies to meet the standards. This includes answering questions such as when and to what extent unsafe code is acceptable within larger software systems, and guidelines on structuring such unsafe dependencies to support compositional reasoning about safety.

Google's commitment

At Google, we're not just advocating for standardization and a memory-safe future, we're actively working to build it.

We are collaborating with industry and academic partners to develop potential standards, and our joint authorship of the recent CACM call-to-action marks an important first step in this process. In addition, as outlined in our Secure by Design whitepaper and in our memory safety strategy, we are deeply committed to building security into the foundation of our products and services.

This commitment is also reflected in our internal efforts. We are prioritizing memory-safe languages, and have already seen significant reductions in vulnerabilities by adopting languages like Rust in combination with existing, wide-spread usage of Java, Kotlin, and Go where performance constraints permit. We recognize that a complete transition to those languages will take time. That's why we're also investing in techniques to improve the safety of our existing C++ codebase by design, such as deploying hardened libc++.

Let's build a memory-safe future together

This effort isn't about picking winners or dictating solutions. It's about creating a level playing field, empowering informed decision-making, and driving a virtuous cycle of security improvement. It's about enabling a future where:

• Developers and vendors can confidently build more secure systems, knowing their efforts can be objectively assessed.

• Businesses can procure memory-safe products with assurance, reducing their risk and protecting their customers.

• Governments can effectively protect critical infrastructure and incentivize the adoption of secure-by-design practices.

• Consumers are empowered to make decisions about the services they rely on and the devices they use with confidence – knowing the security of each option was assessed against a common framework. 

The journey towards memory safety requires a collective commitment to standardization. We need to build a future where memory safety is not an afterthought but a foundational principle, a future where the next generation inherits a digital world that is secure by design.

Acknowledgments

We'd like to thank our CACM article co-authors for their invaluable contributions: Robert N. M. Watson, John Baldwin, Tony Chen, David Chisnall, Jessica Clarke, Brooks Davis, Nathaniel Wesley Filardo, Brett Gutstein, Graeme Jenkinson, Christoph Kern, Alfredo Mazzinghi, Simon W. Moore, Peter G. Neumann, Hamed Okhravi, Peter Sewell, Laurence Tratt, Hugo Vincent, and Konrad Witaszczyk, as well as many others.