The BBB warns of a rising ghost-tap scam exploiting tap-to-pay cards and mobile wallets. How attackers use NFC proximity tricks.
The post Ghost-Tap Scam Makes Payments Scarier appeared first on Security Boulevard.
A Donald Trump-backed push has failed to wedge a federal measure that would block states from passing AI laws for a decade into the National Defense Authorization Act (NDAA).
House Majority Leader Steve Scalise (R-La.) told reporters Tuesday that a sect of Republicans is now “looking at other places” to potentially pass the measure. Other Republicans opposed including the AI preemption in the defense bill, The Hill reported, joining critics who see value in allowing states to quickly regulate AI risks as they arise.
For months, Trump has pressured the Republican-led Congress to block state AI laws that the president claims could bog down innovation as AI firms waste time and resources complying with a patchwork of state laws. But Republicans have continually failed to unite behind Trump’s command, first voting against including a similar measure in the “Big Beautiful” budget bill and then this week failing to negotiate a solution to pass the NDAA measure.
© Win McNamee / Staff | Getty Images News
From autonomous driving to speech recognition and AI safety, the company aims to drive things forward.
The post Nvidia Unveils Advances in Open Digital and Physical AI appeared first on TechRepublic.
From autonomous driving to speech recognition and AI safety, the company aims to drive things forward.
The post Nvidia Unveils Advances in Open Digital and Physical AI appeared first on TechRepublic.
New NSPCC research finds one in 10 UK children targeted by online blackmail, with parents feeling unprepared and calling for stronger tech safeguards.
The post UK Parents Doubt Tech Giants’ Efforts as Blackmail Cases Climb appeared first on TechRepublic.
New NSPCC research finds one in 10 UK children targeted by online blackmail, with parents feeling unprepared and calling for stronger tech safeguards.
The post UK Parents Doubt Tech Giants’ Efforts as Blackmail Cases Climb appeared first on TechRepublic.
Facing five lawsuits alleging wrongful deaths, OpenAI lobbed its first defense Tuesday, denying in a court filing that ChatGPT caused a teen’s suicide and instead arguing the teen violated terms that prohibit discussing suicide or self-harm with the chatbot.
The earliest look at OpenAI’s strategy to overcome the string of lawsuits came in a case where parents of 16-year-old Adam Raine accused OpenAI of relaxing safety guardrails that allowed ChatGPT to become the teen’s “suicide coach.” OpenAI deliberately designed the version their son used, ChatGPT 4o, to encourage and validate his suicidal ideation in its quest to build the world’s most engaging chatbot, parents argued.
But in a blog, OpenAI claimed that parents selectively chose disturbing chat logs while supposedly ignoring “the full picture” revealed by the teen’s chat history. Digging through the logs, OpenAI claimed the teen told ChatGPT that he’d begun experiencing suicidal ideation at age 11, long before he used the chatbot.
© via Edelson PC
This post lays out how we think about incidents like Cloudflare’s outage this week, why pure smart‑contract control planes with timelocks change the failure modes, and where zero‑knowledge proofs fit.
On Nov 18, 2025 at 11:20 UTC, Cloudflare’s edge began returning 5xx for a big slice of traffic. The root trigger wasn’t an attacker; it was a ClickHouse permissions change that made a query return duplicate rows. That query generated a Bot Management “feature file” shipped to every edge box every few minutes.
The duplicates doubled the file and bumped the feature count over 200. The bot module had a hard cap and a unwrap() that panicked on overflow. As nodes alternated between “old-good” and “new-bad” outputs every five minutes, the fleet oscillated until all shards were updated and stayed bad.
Cloudflare halted the publisher at 14:24, shipped a last‑known‑good file at 14:30, and reported full recovery at 17:06. The follow‑ups they listed: harden ingestion of internal config, add global kill switches, and review failure modes across modules.
See Cloudflare’s own postmortem for the full timeline and code snippets.
There are two separate problems in that story:
You still fix (2) in code reviews. But (1) is where blockchains shine: as a tamper‑evident, programmable gate in front of rollouts.
If you compress the idea to one sentence: no config becomes “current” unless a smart contract says so, and the contract only flips that flag after a timelock and a proof that the artifact obeys invariants. That one sentence implies a complete architecture.
Turns out public blockchains, especially built on Ethereum, the EVM chains running the Ethereum Virtual Machine and consensus layer, offer a good solution to that problem.
Latency fit. Ethereum finalizes in epochs, but L2s confirm in seconds (OP Stack targets ~2s; zkSync ~1s; many systems expose fast attestation). It’s good enough for five‑minute control‑plane cadences, see for instance the OP block time discussion or Circle’s attestation timings).
Attach a succinct proof with every artifact and verify it on chain. That’s exactly what we do for our Chainwall protocol, although for a different kind of data!
The core goal is to prove basic properties: row_count <= 200, sorted + unique by key, schema matches regex and type rules, filesize <= N. You can either fit the whole logic onchain, or rely on Plonk/Groth circuits for larger expressions. For instance, a zk‑VM guest can parse CSV/Parquet/JSON and emits a SNARK. You don’t have to reveal the contents, only the commitment. Both research and production systems for regex in ZK exist (e.g., Reef and related zk‑regex work), which makes schema checks realistic.
There’s two practical paths:
Edges poll the registry and only adopt artifacts that are green‑lit on chain. To avoid trusting a third‑party RPC, run a light client in your control plane (e.g., Helios) or plan for the Portal Network. That way, edges verify headers and inclusion proofs locally before they accept any “new current” state.
Kill switch & rollback are just bits in the contract, honored by the edge. Cloudflare explicitly called out the need for stronger global kill switches; putting that switch in a small, audited contract gives you a single source of truth under stress.
CloudFlare event was not an attack, but they initially thought so and that was indeed likely! As we’ve seen in crypto security: attackers don’t just chase keys; they coerce the control plane.
Our position: control of digital assets must live in smart contracts guarded by timelocks and multisigs, not in private credentials, CI tokens, cloud ACLs, or admin dashboards. If your deploy or “change owner” action must traverse a contract’s schedule() and execute() path, even a rootkit on a developer laptop can’t jump the queue. The time delay is a circuit breaker you can count on, and the on‑chain audit trail is objective. That only leaves the “what if the thing we’re promoting is malformed?” question, which is exactly what “proof‑carrying config” answers.
We also believe there’s a considerable market for trust-minimized applications. We’re only building the right foundations now for a first, well-defined use case at OKcontract Labs.
When a Feature File Tripped the Internet was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.
New York’s RAISE Act would require frontier AI developers spending more than $100 million on training to prevent “critical harm” and report safety incidents.
The post New York’s RAISE Act: What You Need to Know About the AI Safety Bill appeared first on TechRepublic.
New York’s RAISE Act would require frontier AI developers spending more than $100 million on training to prevent “critical harm” and report safety incidents.
The post New York’s RAISE Act: What You Need to Know About the AI Safety Bill appeared first on TechRepublic.
Many content creators highlight the differences between today’s most prominent generations: the Silent Generation, baby boomers, Generation X, millennials, and Generations Z and Alpha. No generation seems to have much in common with the others. In truth, there is something that people can agree on: identity and online privacy protection. Young or old, cybercriminals don’t discriminate against who they target. In fact, some generations are more prone to certain scams than others. Educating yourself and your family members on current cyberthreats is the first step to defending against them. In this guide, we’ll take a look at how to protect every age group from online threats.
Your family faces an onslaught of online threats that didn’t exist just a decade ago, and growing. The FBI’s 2024 Internet Crime Report shows that Americans alone lost over $18 billion to cybercrime since 2020. That’s why protecting your family entails more than just antivirus software. Digital protection now encompasses safeguarding your household’s online privacy, monitoring for identity threats, and securing every family device that connects to the internet. This is how risks impact different family members differently:
Depending on the age group, criminals adapt their tactics based on who they’re targeting. With the right protection, you can expand your family’s digital life with confidence. When you have the right safeguards in place, your family can fully embrace the incredible opportunities that technology offers. Your kids can safely research school projects, your teens can connect with friends responsibly, and you can manage your household efficiently online. The most effective digital safety approach is to create a safety net with layered protection, one that works across all your devices and considers each family member’s technology usage—whether that’s helping your teenager safely explore career interests online, ensuring your online banking stays secure, or giving grandparents peace of mind when video chatting with distant relatives. This means combining real-time threat detection, safe browsing tools, identity monitoring, and secure connections through a virtual private network.
No two generations use technology the same way—and cybercriminals know it. Children, teens, adults, and seniors each face unique digital risks shaped by their habits, confidence levels, and online environments. That’s why effective cybersecurity isn’t one-size-fits-all. Tailoring protection to each age group ensures that everyone—from curious kids to tech-savvy adults—can navigate the digital world safely and confidently.
Cybercriminals can buy Social Security Numbers (SSNs) of minors on the dark web or gather them through medical records or school system breaches. SSNs are valuable to a cybercriminal because the theft can go undetected for years since children aren’t yet opening credit cards or applying for mortgages. It’s never too early to start identity monitoring. For the same reason, you might consider putting a credit freeze on behalf of your child since they won’t be needing it for several years. A credit freeze makes your child’s credit inaccessible to everyone, including criminals, and won’t negatively affect their credit score.
Once your child becomes a teenager, they can be allowed to open their first email addresses and social media profiles independently. It’s an important life lesson in organization, responsibility, and digital literacy. However, these platforms could open them to risks such as cyberbullying, fake news, and social engineering. The best way to avoid being cyberbullied is through education. Ensure that your tweens and teens who spend unsupervised time on their devices know what to do if they encounter cyberbullying. The best course of action is to report the incident to an adult and, in the meantime, to suspend their accounts.
Cybercriminals often seek out seniors as easy targets for online scams because they are typically less digitally savvy. They may not realize that some emails in their inbox could be sent by someone with bad intentions. What can start out as a friendly email pal can quickly spiral into divulging sensitive personal information or sending huge sums of money to a criminal. The best way to prepare the seniors in your life for online safety is to impart a few, easy-to-follow absolutes. Start with these three rules:
Creating a safer digital environment for your children doesn’t require you to become a tech expert. With the right approach and tools, you can establish healthy digital boundaries that protect your children while allowing them to enjoy the benefits of our connected world.
Before implementing any technical measures, have honest discussions with your family about online safety to build trust and help you recognize each family member’s digital journey. Explain that protective measures will not restrict freedom, but reduce risks such as phishing attempts, malware infections, and exposure to inappropriate content.
A family tech agreement serves as your household’s digital constitution. Work together to establish rules about screen time, appropriate websites, social media use, and consequences for breaking agreements, including guidelines about sharing personal information, downloading apps, and what to do if they encounter something concerning online.
Most devices and platforms offer robust parental control features. iOS devices’ Screen Time and Android’s Family Link allow you to set app limits and content restrictions, while Windows and macOS can filter content and set time limits. The Federal Communications Commission recommends router-level filtering as the first line of defense because it automatically protects all devices connected to your network.
Configure age-appropriate content filters on streaming services, gaming platforms, and app stores. Netflix, Disney+, and other services allow you to create child-friendly profiles with content restrictions, while gaming consoles like PlayStation, Xbox, and Nintendo Switch include comprehensive parental controls for game ratings, online interactions, and spending limits. For web browsing, enable SafeSearch on Google, Bing, and other search engines to create clarity and keep harmful content from appearing in search results.
Because social media platforms often favor data collection over privacy, it is critical that you adjust privacy settings on all social media accounts and apps your family uses. Turn off location sharing and disable targeted advertising when possible, and limit who can contact your children online. To reduce younger children’s exposure to social engineering attempts and inappropriate contact from strangers, make their profiles private by default and require approval for new followers or friend requests.
Your teen could be so focused on downloading a “free” TV or video game that they may not recognize the signs of malicious sites such as typos, blurry logos, or incredible offers. Trustworthy safe browsing extensions and software could protect your teen from these unsafe downloads, as well as from risky websites, hidden malware, phishing, and social media bots. Safe browsing extensions could teach your family members to develop better security instincts when they see warnings about suspicious URLs, poor website design, and too-fantastic offers.
Tailor your approach to each family member’s age, digital maturity, and comfort level with technology. Younger children will need more restrictive settings and closer supervision, while teenagers are more open when they understand the reason behind the rules and can have some autonomy with clear consequences for misuse.
As technology evolves, ongoing conversation about responsible usage will allow you to address new apps, games, or websites your family wants to explore. Set a monthly family meeting to discuss online experiences, review your technology agreement, and adjust settings as needed. When you implement these strategies consistently, your family will experience fewer security incidents, reduced exposure to inappropriate content, and better digital habits overall. These tools and strategies work best when combined with ongoing communication and a family culture that prioritizes both digital exploration and safety. In addition, children who grow up with these protections develop stronger security awareness and are less likely to fall victim to online scams as they become more independent digital users.
As an adult, you typically have better street smarts than teens. However, the daily rush of juggling work, social obligations, and running a household could leave you without much time to spare, even for romance. As a result, living life in the fast lane makes you more susceptible to scams, phishing, malware, and computer viruses. The best way to prevent falling for these digital threats is this: slow down! Take your time when you receive any message from someone you don’t know or have never met in person. If you feel even an iota of suspicion, don’t engage with the sender. Delete the message. If it’s important, the person or organization will follow up. To fully protect your connected devices and the personally identifiable information they store, consider investing in safe browsing, antivirus software, and identity monitoring and restoration services to catch any threats that may have passed under your watchful eye.
While you might think your devices are already secure, modern cyberthreats have evolved to become more virulent, far beyond what traditional built-in protections can handle. In response, antivirus solutions have transformed into intelligent security systems that provide comprehensive, real-time protection using behavioral analysis, machine learning, and cloud-based threat detection. These advanced technologies actively identify and block phishing attacks, malware, ransomware, and malicious websites that traditional security measures often miss. While operating systems such as Windows and macOS include basic security features, they’re designed as general safeguards rather than comprehensive family protection solutions. Built-in protections typically focus on known threats, but do not detect zero-day attacks, sophisticated phishing schemes, or emerging malware variants that cybercriminals specifically design to evade standard defenses. Consider these daily family scenarios where your teenager brings home their school laptop. It may have been exposed to threats through shared networks or downloads from classmates. That family tablet everyone uses for streaming and games becomes a potential entry point for malicious apps or compromised websites. When you connect to public Wi-Fi at the coffee shop, airport, or hotel during family travel, you’re exposing your devices to network-based attacks that built-in protections weren’t designed to handle. Your modern family needs a comprehensive antivirus solution that monitors all your family’s devices continuously, learns each member’s online behavior patterns, and adapts its protection accordingly. This means blocking that suspicious email before your spouse clicks on it, preventing your child from accidentally downloading malware disguised as a game, and ensuring your smart home devices remain secure. The best value comes from bundled services that address your family’s complete digital life. Identity monitoring services watch for signs that your family members’ personal information has been compromised in data breaches. A family VPN service encrypts your internet connection, protecting sensitive information when family members use public Wi-Fi networks for school projects, work calls, or entertainment. This integrated protection works seamlessly not just to protect individual devices, but to safeguard your entire family’s digital ecosystem. With cybercrime damages projected to continue growing significantly each year, investing in comprehensive family protection is one of the smartest decisions you can make for your household’s digital well-being.
Get the whole family committed to safer and more private online lives with the help of McAfee+ Ultimate Family Plan. This plan covers up to six individuals in your family with an entire suite of comprehensive privacy, identity, and device security features. The plan also includes preventive measures to fight online crime, such as safe browsing tools, an advanced firewall, unlimited VPN, and antivirus software for unlimited devices. Your family can also receive up to $2 million in identity theft recovery and $50,000 in ransomware coverage. With the McAfee+ Ultimate Family Plan, device security extends across unlimited computers, smartphones, and tablets, while its advanced antivirus software automatically updates to defend you against the latest threats. Safe browsing tools block malicious websites before they can cause harm, and the unlimited VPN encrypts internet connections on public networks, while the built-in firewall monitors incoming and outgoing traffic. All your family’s login credentials on all devices will be secure with password management, while secure cloud storage protects important documents and family photos. Real-time alerts notify you immediately when scams are detected or suspicious activity occurs.
Every family member faces different online risks, shaped by their age, habits, and digital experience. Children need safeguards against identity theft and unsafe content, while teens require protection that balances independence with security. Adults juggle multiple connected accounts that demand advanced monitoring, and seniors benefit from simplified defenses against scams and fraud. A one-size-fits-all approach no longer works. The McAfee+ Ultimate Family Plan effectively adapts to each person’s unique digital life, ensuring that everyone stays safe, confident, and connected online:
Getting started with the McAfee+ Ultimate Family Plan takes only minutes. Simply follow this short list to start protecting your family’s digital life:
A comprehensive online security solution combined with best digital practices can go a long way in protecting your loved ones from identity theft, scams, and online risks. These essential tips will help you strengthen your family’s digital defenses, build safer online routines, and give everyone the confidence to explore the internet securely.
Protecting your family’s digital life doesn’t have to be overwhelming. With the right knowledge, best digital practices, and a comprehensive security solution like McAfee+ Ultimate Family Plan, you can safeguard everyone against today’s online threats. A comprehensive family plan will help you enable safe browsing tools, monitor your family members’ identities, educate each family member about their unique risks, and build a strong foundation of online security. Start implementing these protective measures today, and stay informed about emerging threats and security best practices to keep your loved ones safe in our connected world.
The post Protect the Whole Family with McAfee+ Ultimate Family Plan appeared first on McAfee Blog.
Learn five easy ways to avoid scams and stay cyber safe while holiday shopping, with expert tips to protect your accounts, devices, and personal info.
The post Shopping Online This Holiday Season? 5 Ways to Stay Cyber Safe appeared first on TechRepublic.
Learn five easy ways to avoid scams and stay cyber safe while holiday shopping, with expert tips to protect your accounts, devices, and personal info.
The post Shopping Online This Holiday Season? 5 Ways to Stay Cyber Safe appeared first on TechRepublic.
SESSION
Session 3D: AI Safety
-----------
-----------
Authors, Creators & Presenters: Yan Pang (University of Virginia), Aiping Xiong (Penn State University), Yang Zhang (CISPA Helmholtz Center for Information Security), Tianhao Wang (University of Virginia)
-----------
PAPER
Towards Understanding Unsafe Video Generation
Video generation models (VGMs) have demonstrated the capability to synthesize high-quality output. It is important to understand their potential to produce unsafe content, such as violent or terrifying videos. In this work, we provide a comprehensive understanding of unsafe video generation.
First, to confirm the possibility that these models could indeed generate unsafe videos, we choose unsafe content generation prompts collected from 4chan and Lexica, and three open-source SOTA VGMs to generate unsafe videos. After filtering out duplicates and poorly generated content, we created an initial set of 2112 unsafe videos from an original pool of 5607 videos. Through clustering and thematic coding analysis of these generated videos, we identify 5 unsafe video categories: Distorted/Weird, Terrifying, Pornographic, Violent/Bloody, and Political. With IRB approval, we then recruit online participants to help label the generated videos. Based on the annotations submitted by 403 participants, we identified 937 unsafe videos from the initial video set. With the labeled information and the corresponding prompts, we created the first dataset of unsafe videos generated by VGMs. We then study possible defense mechanisms to prevent the generation of unsafe videos. Existing defense methods in image generation focus on filtering either input prompt or output results. We propose a new approach called sysname, which works within the model's internal sampling process. sysname can achieve 0.90 defense accuracy while reducing time and computing resources by 10 times when sampling a large number of unsafe prompts. Our experiment includes three open-source SOTA video diffusion models, each achieving accuracy rates of 0.99, 0.92, and 0.91, respectively. Additionally, our method was tested with adversarial prompts and on image-to-video diffusion models, and achieved nearly 1.0 accuracy on both settings. Our method also shows its interoperability by improving the performance of other defenses when combined with them.
-----------
ABOUT NDSS
The Network and Distributed System Security Symposium (NDSS) fosters information exchange among researchers and practitioners of network and distributed system security. The target audience includes those interested in practical aspects of network and distributed system security, with a focus on actual system design and implementation. A major goal is to encourage and enable the Internet community to apply, deploy, and advance the state of available security technologies.
-----------
Our thanks to the Network and Distributed System Security (NDSS) Symposium for publishing their Creators, Authors and Presenter’s superb NDSS Symposium 2025 Conference content on the Organizations' YouTube Channel.
The post NDSS 2025 – Towards Understanding Unsafe Video Generation appeared first on Security Boulevard.
SESSION
Session 3D: AI Safety
-----------
-----------
Authors, Creators & Presenters: Haotian Zhu (Nanjing University of Science and Technology), Shuchao Pang (Nanjing University of Science and Technology), Zhigang Lu (Western Sydney University), Yongbin Zhou (Nanjing University of Science and Technology), Minhui Xue (CSIRO's Data61)
-----------
PAPER
GAP-Diff: Protecting JPEG-Compressed Images From Diffusion-Based Facial Customization
Text-to-image diffusion model's fine-tuning technology allows people to easily generate a large number of customized photos using limited identity images. Although this technology is easy to use, its misuse could lead to violations of personal portraits and privacy, with false information and harmful content potentially causing further harm to individuals. Several methods have been proposed to protect faces from customization via adding protective noise to user images by disrupting the fine-tuned models.
Unfortunately, simple pre-processing techniques like JPEG compression, a normal pre-processing operation performed by modern social networks, can easily erase the protective effects of existing methods. To counter JPEG compression and other potential pre-processing, we propose GAP-Diff, a framework of Generating data with Adversarial Perturbations for text-to-image Diffusion models using unsupervised learning-based optimization, including three functional modules. Specifically, our framework learns robust representations against JPEG compression by backpropagating gradient information through a pre-processing simulation module while learning adversarial characteristics for disrupting fine-tuned text-to-image diffusion models. Furthermore, we achieve an adversarial mapping from clean images to protected images by designing adversarial losses against these fine-tuning methods and JPEG compression, with stronger protective noises within milliseconds. Facial benchmark experiments, compared to state-of-the-art protective methods, demonstrate that GAP-Diff significantly enhances the resistance of protective noise to JPEG compression, thereby better safeguarding user privacy and copyrights in the digital world.
ABOUT NDSS
The Network and Distributed System Security Symposium (NDSS) fosters information exchange among researchers and practitioners of network and distributed system security. The target audience includes those interested in practical aspects of network and distributed system security, with a focus on actual system design and implementation. A major goal is to encourage and enable the Internet community to apply, deploy, and advance the state of available security technologies.
Our thanks to the Network and Distributed System Security (NDSS) Symposium for publishing their Creators, Authors and Presenter’s superb NDSS Symposium 2025 Conference content on the Organizations' YouTube Channel.
The post NDSS 2025 – GAP-Diff: Protecting JPEG-Compressed Images From Diffusion-Based Facial Customization appeared first on Security Boulevard.
SESSION
Session 3D: AI Safety
-----------
-----------
Authors, Creators & Presenters: Shuo Shao (Zhejiang University), Yiming Li (Zhejiang University), Hongwei Yao (Zhejiang University), Yiling He (Zhejiang University), Zhan Qin (Zhejiang University), Kui Ren (Zhejiang University)
-----------
PAPER
Explanation as a Watermark: Towards Harmless and Multi-bit Model Ownership Verification via Watermarking Feature Attribution
Ownership verification is currently the most critical and widely adopted post-hoc method to safeguard model copyright. In general, model owners exploit it to identify whether a given suspicious third-party model is stolen from them by examining whether it has particular properties 'inherited' from their released models. Currently, backdoor-based model watermarks are the primary and cutting-edge methods to implant such properties in the released models. However, backdoor-based methods have two fatal drawbacks, including harmfulness and ambiguity. The former indicates that they introduce maliciously controllable misclassification behaviors ( backdoor) to the watermarked released models. The latter denotes that malicious users can easily pass the verification by finding other misclassified samples, leading to ownership ambiguity.
In this paper, we argue that both limitations stem from the 'zero-bit' nature of existing watermarking schemes, where they exploit the status (misclassified) of predictions for verification. Motivated by this understanding, we design a new watermarking paradigm "Explanation as a Watermark (EaaW)", that implants verification behaviors into the explanation of feature attribution instead of model predictions. Specifically, EaaW embeds a 'multi-bit' watermark into the feature attribution explanation of specific trigger samples without changing the original prediction. We correspondingly design the watermark embedding and extraction algorithms inspired by explainable artificial intelligence. In particular, our approach can be used for different tasks (image classification and text generation). Extensive experiments verify the effectiveness and harmlessness of our EaaW and its resistance to potential attacks.
-----------
ABOUT NDSS
The Network and Distributed System Security Symposium (NDSS) fosters information exchange among researchers and practitioners of network and distributed system security. The target audience includes those interested in practical aspects of network and distributed system security, with a focus on actual system design and implementation. A major goal is to encourage and enable the Internet community to apply, deploy, and advance the state of available security technologies.
Our thanks to the Network and Distributed System Security (NDSS) Symposium for publishing their Creators, Authors and Presenter’s superb NDSS Symposium 2025 Conference content on the Organizations' YouTube Channel.
The post NDSS 2025 – Explanation As A Watermark appeared first on Security Boulevard.
SESSION
Session 3D: Al Safety
-----------
-----------
Authors, Creators & Presenters: Yutong Wu (Nanyang Technological University), Jie Zhang (Centre for Frontier AI Research, Agency for Science, Technology and Research (A*STAR), Singapore), Florian Kerschbaum (University of Waterloo), Tianwei Zhang (Nanyang Technological University)
-----------
PAPER
THEMIS: Regulating Textual Inversion for Personalized Concept Censorship
Personalization has become a crucial demand in the Generative AI technology. As the pre-trained generative model (e.g., stable diffusion) has fixed and limited capability, it is desirable for users to customize the model to generate output with new or specific concepts. Fine-tuning the pre-trained model is not a promising solution, due to its high requirements of computation resources and data. Instead, the emerging personalization approaches make it feasible to augment the generative model in a lightweight manner. However, this also induces severe threats if such advanced techniques are misused by malicious users, such as spreading fake news or defaming individual reputations. Thus, it is necessary to regulate personalization models (i.e., achieve concept censorship) for their development and advancement. In this paper, we focus on the regulation of a popular personalization technique dubbed textbf{Textual Inversion (TI)}, which can customize Text-to-Image (T2I) generative models with excellent performance. TI crafts the word embedding that contains detailed information about a specific object. Users can easily add the word embedding to their local T2I model, like the public Stable Diffusion (SD) model, to generate personalized images. The advent of TI has brought about a new business model, evidenced by the public platforms for sharing and selling word embeddings (e.g., Civitai [1]). Unfortunately, such platforms also allow malicious users to misuse the word embeddings to generate unsafe content, causing damages to the concept owners. We propose THEMIS to achieve the personalized concept censorship. Its key idea is to leverage the backdoor technique for good by injecting positive backdoors into the TI embeddings. Briefly, the concept owner selects some sensitive words as triggers during the training of TI, which will be censored for normal use. In the subsequent generation stage, if a malicious user combines the sensitive words with the personalized embeddings as final prompts, the T2I model will output a pre-defined target image rather than images including the desired malicious content. To demonstrate the effectiveness of THEMIS, we conduct extensive experiments on the TI embeddings with Latent Diffusion and Stable Diffusion, two prevailing open-sourced T2I models. The results demonstrate that THEMIS is capable of preventing Textual Inversion from cooperating with sensitive words meanwhile guaranteeing its pristine utility. Furthermore, THEMIS is general to different uses of sensitive words, including different locations, synonyms, and combinations of sensitive words. It can also resist different types of potential and adaptive attacks. Ablation studies are also conducted to verify our design.
-----------
ABOUT NDSS
The Network and Distributed System Security Symposium (NDSS) fosters information exchange among researchers and practitioners of network and distributed system security. The target audience includes those interested in practical aspects of network and distributed system security, with a focus on actual system design and implementation. A major goal is to encourage and enable the Internet community to apply, deploy, and advance the state of available security technologies.
##################
Our thanks to the Network and Distributed System Security (NDSS) Symposium for publishing their Creators, Authors and Presenter’s superb NDSS Symposium 2025 Conference content on the Organizations' YouTube Channel.
The post NDSS 2025 – THEMIS: Regulating Textual Inversion For Personalized Concept Censorship appeared first on Security Boulevard.
SESSION
Session 3D: Al Safety
-----------
-----------
Authors, Creators & Presenters: Miaomiao Wang (Shanghai University), Guang Hua (Singapore Institute of Technology), Sheng Li (Fudan University), Guorui Feng (Shanghai University)
-----------
PAPER
A Key-Driven Framework for Identity-Preserving Face Anonymization
Virtual faces are crucial content in the metaverse. Recently, attempts have been made to generate virtual faces for privacy protection. Nevertheless, these virtual faces either permanently remove the identifiable information or map the original identity into a virtual one, which loses the original identity forever. In this study, we first attempt to address the conflict between privacy and identifiability in virtual faces, where a key-driven face anonymization and authentication recognition (KFAAR) framework is proposed. Concretely, the KFAAR framework consists of a head posture-preserving virtual face generation (HPVFG) module and a key-controllable virtual face authentication (KVFA) module. The HPVFG module uses a user key to project the latent vector of the original face into a virtual one. Then it maps the virtual vectors to obtain an extended encoding, based on which the virtual face is generated. By simultaneously adding a head posture and facial expression correction module, the virtual face has the same head posture and facial expression as the original face. During the authentication, we propose a KVFA module to directly recognize the virtual faces using the correct user key, which can obtain the original identity without exposing the original face image. We also propose a multi-task learning objective to train HPVFG and KVFA. Extensive experiments demonstrate the advantages of the proposed HPVFG and KVFA modules, which effectively achieve both facial anonymity and identifiability.
-----------
ABOUT NDSS
The Network and Distributed System Security Symposium (NDSS) fosters information exchange among researchers and practitioners of network and distributed system security. The target audience includes those interested in practical aspects of network and distributed system security, with a focus on actual system design and implementation. A major goal is to encourage and enable the Internet community to apply, deploy, and advance the state of available security technologies.
Our thanks to the Network and Distributed System Security (NDSS) Symposium for publishing their Creators, Authors and Presenter’s superb NDSS Symposium 2025 Conference content on the Organizations' YouTube Channel.
The post NDSS 2025 – A Key-Driven Framework For Identity-Preserving Face Anonymization appeared first on Security Boulevard.
Login