All items
All tracked items across vulnerabilities, news, research, incidents, and regulatory updates.
All tracked items across vulnerabilities, news, research, incidents, and regulatory updates.
48 items
OpenAI will acquire Pierre Computer Company, Inc., the company behind Code Storage and open-source developer tools including Trees and Diffs, which Codex already uses to navigate files and review changes. Pierre's open-source projects will remain open-source, and Code Storage stays available to customers after closing while its features are brought into Codex. The deal is subject to customary closing conditions, including regulatory approvals, and until closing the two companies remain separate.
Anthropic's SEC filing for its initial public offering devotes 93 pages to risk factors, far more than rival SpaceX's 38 pages. The risks range from losing key staff and regulatory obstacles to AI-enabled cyberattacks, misinformation, and autonomous systems causing harm at scale. Anthropic presents the IPO as a step toward mitigating the risks it believes AI poses.
This review examines recent developments in privacy-preserving federated learning for healthcare, covering medical imaging, EHR analytics, genomics and wearable monitoring, with work published between 2024 and early 2026. It analyses challenges including non-IID data, communication overhead, gradient leakage, model poisoning and membership inference attacks, along with regulatory and interoperability concerns. The paper was published in Discover Applied Sciences on 2026-10-16.
Manus, an AI agent startup, raised more than $500 million in its first funding round since Meta abandoned its acquisition of the company. The round was led by Boyu Capital and IDG Capital, with follow-on investment from existing shareholders Tencent, HSG and ZhenFund, and the post-funding valuation was not disclosed. Chinese regulators had blocked Meta's roughly $2 billion deal, and analysts say Manus must now prove scale, profitability and regulatory alignment.
After Anthropic CEO Dario Amodei urged leading AI labs to "pace the frontier," Sam Altman and Elon Musk publicly backed his call for a pause, and former Anthropic researcher Jacob Coxon and former DeepMind safety researcher Josh Engels resigned citing safety concerns. The article argues that agentic cybersecurity threats deserve more attention than the debate over regulatory capture and frontier labs' commitment to responsible AI, and that the central challenge is agreeing how to slow AI development without ceding economic and security advantages.
An expert panel at ECIS 2025 in Amman, Jordan, examined real-time data sharing frameworks (RTDSFs), which support low-latency, cross-organisational data flows in finance, healthcare and energy. Through interpretive analysis of the panel transcript and polling of eighteen participants, the authors propose four pillars for Information Systems research: trust, governance, federated intelligence, and security collaboration under cascade risk. Polling results are presented as indicative only, showing alignment on enforceable governance and regulatory engagement.
Former DOJ antitrust chief Jonathan Kanter discusses AI safety and regulation on the Decoder podcast. The source describes researchers at Anthropic and Google DeepMind quitting over safety concerns, and CEOs asking for antitrust exemptions to coordinate on safety. Critics have accused these companies of seeking regulatory capture or forming a cartel.
OpenAI's nonprofit parent, the OpenAI Foundation, announced it will fund a proposal by policy analyst Ruxandra Teslo to obtain data from failed biotech companies through bankruptcy proceedings, paying to create "high-quality scientific datasets" to support medical AI. Teslo argued that bids at bankruptcy sales could yield regulatory filings, manufacturing strategies and safety data, which she called "biotech's lost archive." The same newsletter also covers an economic analysis of hyperscaler AI spending.
The OpenAI Foundation, the nonprofit parent of OpenAI, launched Public Data for Health to fund the creation of high-quality scientific datasets for AI in biology and medicine. Its first round of data grants includes $40 million for a University of North Carolina, Chapel Hill program collecting data on novel cancer vaccines, and $500,000 for Ruxandra Teslo's proposal, pursued by 1Day Sooner, to acquire regulatory filings and safety data from failed biotech companies through bankruptcy proceedings.
Cybersecurity professionals moving toward CISO or executive roles need skills beyond technical expertise, such as translating technical risk into business priorities. A study of CISO job postings found that employers value STEM or business education, vendor-neutral certification, strong communication skills, and knowledge of regulatory frameworks, while placing little emphasis on platform mastery, coding ability, or security clearances.
A bioengineer's lab at César de la Fuente's group uses deep-learning models trained on biological sequences, along with ChatGPT and Codex, to search genome and protein datasets for antimicrobial candidates. The article says this can cut the initial candidate search from years to hours. Candidates still require laboratory validation, toxicity and resistance testing, and regulatory review before becoming drugs.
This regulatory guidance advises that AI agents always run within a sandboxed environment that controls what resources they can reach locally and over a network. It recommends defining the sandbox boundaries across execution, network, compute, credentials and data, and denying network traffic by default, using allowlists or protocol-aware proxies with manual approval. It also warns that agents may discover and exploit configuration issues or vulnerabilities in technical controls, leading to a sandbox escape, so it calls for multiple isolation layers and regular validation.
A series of advances in China's artificial intelligence, chip manufacturing and robotics has unsettled financial markets and divided US tech leaders. The Trump administration is reportedly struggling to respond. Silicon Valley has long cited Chinese tech growth to argue against regulatory oversight, but recent progress has pushed US tech CEOs into open disagreement over how to address Chinese-made products.
The U.S. Department of Commerce has reportedly cleared OpenAI to proceed with a broad release of GPT-5.6, according to Axios citing an unnamed source. OpenAI expects the rollout as early as this week after further testing and meetings with government officials. The approval follows a phased launch of GPT-5.6 Sol, Terra and Luna, initially limited to a small group of trusted partners.
Insignary, Inc. announced it was named a Sample Vendor for Reachability Analysis in the Gartner Hype Cycle for Secure Software Engineering, 2026, and has been cited in four Gartner research reports. Its Insignary Clarity platform builds a Software Bill of Materials (SBOM) by scanning compiled binaries and source, identifying open-source components that never appear in package manifests.
KNOX is presented as the first regulatory sandbox for generative models, built by running AI workloads inside TEEs and demonstrated on agentic workloads. For compliance, it derives an attested policy manifest per user prompt and uses an attested reference monitor to allow only actions that match it, recording every action and verdict in an attested execution graph. The authors report that KNOX meets all four requirements, detects deviations such as those caused by prompt injection, and incurs less than 1% overhead.
This critical review examines privacy risks from conversational LLM agents across five areas: data leakage and memorization, adversarial extraction and prompt injection, inference and re-identification, surveillance and profiling, and regulatory governance breakdown. It finds that technical mitigations such as differential privacy, federated learning and machine unlearning work to some extent in controlled settings but fail to fully protect conversational deployments, and it highlights risks that machine unlearning can leave traces exploitable by adversarial reconstruction.
Fix: The authors support a multi-layered mitigation framework combining Privacy-Preserving Data Publishing (PPDP) principles, technical safeguards, and governance reform.
OpenAlex (peer-reviewed AI security)This paper presents an integrated framework for financial fraud detection that combines ensemble models such as XGBoost, LightGBM, and CatBoost with recurrent and graph neural networks, plus behavioral biometrics, SHAP and LIME explainability, and federated learning. It also examines adversarial attacks on fraud systems and regulatory considerations including the EU AI Act. The paper was published in the International Journal for Research in Applied Science and Engineering Technology (IJRASET) on 2026-09-30.
Fix: Run AI agents within a sandboxed environment that controls and manages what resources can and cannot be communicated with, both locally and over a network. Deny all inbound and outbound network traffic by default and allow only required connections using allowlists, or use protocol- or service-aware proxies that permit connections by exception with manual approval. Restrict each agent to the minimum resources its task needs, use multiple layers of isolation, choose mature sandbox technologies, and regularly validate configurations.
UK NCSCMF2DA is a unified end-to-end framework for detecting and attributing AI-generated images, aimed at robustness to real-world distortions such as social media compression. It combines an Edge Pyramid Fusion ResNet (EPF-ResNet) for pixel-level edge artifacts, a frozen CLIP-ViT for semantic generalization, an MLLM-Guided Quality Refinement Module (MQRM), and a Dual-stream Differential Patch Attribution Network (D2PAN) for generator fingerprints. The authors report state-of-the-art detection of both "friendly" and "ex-regulatory" images with strong cross-dataset generalization across multiple benchmarks.