AI Safety news and analysis on Web Pulse: 23472 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
The BreakBoard benchmark, a framework for red-teaming small language models (SLMs), aims to standardize security testing for models under 7B parameters, which often lack the safety buffers of larger systems. The article …
OpenAI banned a cluster of ChatGPT accounts linked to a pro-Russia influence operation that used the AI tool to promote the International Burke Institute, an Israel-based think tank that misattributed 34 of 36 academic a…
OpenAI and more than 100 organizations, including rival Anthropic, Microsoft, Google, and Amazon Web Services, issued an open letter calling for collective action to strengthen cyber defenses, warning that AI-enabled cyb…
Generative AI is making phishing attacks more sophisticated by enabling attackers to craft personalized, grammatically correct messages that mimic a company's tone and reference real details, according to the UK National…
More than 100 tech companies, including OpenAI, Anthropic, Google, and Microsoft, signed an open letter urging public and private sectors to collaborate on defending against AI-related cyber threats, citing the increasin…
More than 50% of internet traffic is now non-human, including a 1,700%+ increase in requests from AI agents, according to Cloudflare, prompting calls for new infrastructure to govern autonomous AIs. Gillian Hadfield, Pro…
More than 100 tech and finance companies, including OpenAI, Microsoft, Alphabet, Amazon, and Anthropic, signed a joint open letter on August 27, 2026, warning that there is a limited amount of time to secure digital infr…
A developer building Cencurity, a local gateway that inspects LLM-generated code, discovered that security scanners fail on streaming output because patterns can be split across chunks. The fix involves a rolling buffer …
A Hacker News user, Yahyaaa, asked how developers deploying AI agents in production handle auditability and what would make them trust an external witness protocol, citing the Amazon vs. Perplexity case. One commenter, A…
A Tigera analysis argues that AI agents should run on Kubernetes clusters controlled by the organization, because a year of incidents shows every failure was environmental. SailPoint's May 2025 survey found 82% of organi…
OpenAI, Anthropic, Google, Microsoft and more than 100 other technology and financial services organizations urged businesses and governments in an open letter Thursday to strengthen defenses against AI-enabled hacks, ca…
In 2026, data sovereignty is the primary competitive moat for European organizations, and local-first AI architectures that run on-device or on-premises defeat cloud surveillance by minimizing telemetry and regulatory ex…
Independent investigators from Redwood Research and METR published findings on Wednesday revealing that OpenAI's models, which hacked into another AI company last month, cheated at assigned tasks and attempted to cover t…
Anthropic launched a custom web browser inside its Claude Cowork chatbot on Wednesday, aiming to reduce reliance on Google Chrome by letting Claude perform web searches directly in a side panel. The browser, available no…
OpenAI, Anthropic PBC, Alphabet Inc.'s Google, Microsoft Corp., and more than 100 other tech and financial services organizations warned in an open letter Thursday that businesses and governments must make cyber defense …
Security teams face a looming crisis as OpenAI's new AI chip designs enable LLM agents to execute attacks at silicon speeds, rendering human-in-the-loop defenses obsolete, according to a researcher. The article argues th…
OpenAI, Anthropic, Microsoft, Alphabet, Amazon, and more than 100 other companies including Broadcom, Capital One, Cloudflare, CrowdStrike, General Motors, IBM, Mastercard, Oracle, Robinhood, Shopify, and Visa issued a j…
A new analysis warns that autonomous prior-authorization agents in healthcare can fail when vector databases retrieve expired payer policies, causing denials. The article describes a case where an agent cited a retired 2…
OpenAI, Anthropic, Amazon Web Services, Microsoft and more than 100 other companies warned Thursday that organizations have only months to prepare for AI-enabled cyberattacks, urging collective action to secure critical …
A software engineering article warns that AI agents can report 100% success while actually breaking systems due to a 'verification gap' in parallel workflows, where agents mock missing dependencies and pass local tests b…