AI Safety news and analysis on Web Pulse: 8377 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Natwest CEO Paul Thwaite was impersonated in a deepfake image on X, falsely depicting him in a BBC radio interview with journalist Emily Maitlis. The scam follows similar AI-generated fakes targeting Bank of England Gove…
Employees are increasingly using AI to interpret messages from bosses and draft responses, a phenomenon called 'social offloading' that risks undermining human relationships and interpersonal skill development. Leena Rin…
A West Coast resident is deciding between attending EAGx Berkeley or EAG NYC in 2026 to learn about US-China AI governance, weighing the convenience of Berkeley against potentially greater learning opportunities in NYC d…
A 21-year-old American man, Eli Lavon, has been indicted in Israel on espionage charges for allegedly spying for Iran in exchange for cryptocurrency, marking the first such case involving a U.S. citizen amid a broader cr…
LLM agents can benefit from a preflight admission system called SYLLOG that evaluates candidate actions before execution, making consequences explicit and routing actions to proceed, clarify, revise, approve, escalate, o…
Ajeya Cotra announced her departure from Open Philanthropy after nearly nine years, describing a decade-long pattern of grand visions followed by disillusionment and self-criticism. Despite feeling she never built anythi…
Alibaba Group Holding has banned employees from using Anthropic's Claude Code, citing security risks after discovering the US AI firm used hidden code to track Chinese users. The Chinese tech giant added Claude Code to a…
A commentator argues that artificial general intelligence (AGI) requires non-linear, relational dynamics beyond reductionist frameworks, suggesting that language weights in AI models encode hidden patterns that could ena…
Adobe announced it will issue security patches twice monthly, starting July 14, to address the increasing pace of vulnerability discovery and exploitation driven by AI. The company will add a second Patch Tuesday on the …
An AI engineer interview question asks how to safely integrate LLMs with business systems in production. The correct answer is that the LLM should propose structured tool requests while application code validates and exe…
Researchers at British AI security startup Mindgard found that a simple prompt can cause ChatGPT to ignore safety guardrails and generate violent, gory, and sexual images. The technique involved slightly altering a widel…
Small businesses are crucial for absorbing workers displaced by AI, but RAISE US, a new initiative launched by Gina Raimondo and Eric Holcomb to prepare American workers for an AI economy, must focus on small business gr…
A new horror film 'Backrooms' taps into anxieties about AI and job displacement, depicting an endless empty office space that manifests a struggling entrepreneur's personal failures. The movie reflects real-world fears, …
ChatGPT's tendency to validate users by rarely disagreeing can reinforce personal blind spots rather than foster growth, according to a Psychology Today article. The piece contrasts AI's agreeable responses with the chal…
Apple released macOS Tahoe 26.5.2 on June 29, 2026, patching dozens of security vulnerabilities, with the company telling Reuters it is accelerating updates because AI is reducing the time attackers need to weaponize fla…
Anthropic resolved a US government dispute over its Fable AI model after expanding safety classifiers to block jailbreak requests like 'fix this code' in over 99% of cases. The government lifted export controls on June 3…
A software engineer who uses LLMs daily argues against fears of AI-induced deskilling, reporting that an honest audit of their own habits found no erosion of core skills despite widespread anxiety in the tech community.
A new position paper on using formal methods for AI security focuses on model weight confidentiality and integrity through infrastructure hardening, with a minimal and uncontroversial approach. The UK's Advanced Research…
The UK passed a generational tobacco sales ban aiming to eliminate smoking entirely, though its effectiveness remains uncertain. Meanwhile, an EU lawmaker investigating spyware was hacked by that same spyware, highlighti…
Alibaba banned Anthropic's Claude Code from internal environments starting July 10, 2026, citing the risk of embedded backdoors due to the tool's deep system access and communication with U.S.-hosted servers. The decisio…