SafetySecurity Research Finds ~16,000 Supabase Databases Exposing Personal DataSecurity firm UpGuard found roughly 16,000 Supabase-hosted databases that exposed at least some personal data due to misconfiguration.3 min read
SafetyPentagon seeks $30.3M for AI-driven, contactless 'Polygraph+' programThe US Department of Defense has requested $30.3 million over five years to develop “Polygraph+” (aka “Polygraph Next”), a program aiming to combine artificial intelligence and standoff sensing to modernize federal polygraph and credibility assessment tools.5 min read
SafetyMeta prioritizes private processing for Muse assistant as privacy centerpieceMeta is repositioning privacy as a core feature of new AI products, including its viral assistant Muse, offering users options for private processing and even personal virtual machines that Meta cannot access.3 min read
SafetyNvidia CEO Jensen Huang calls existential AI fears a 'distraction', drawing criticismNvidia CEO Jensen Huang has dismissed concerns about AI posing existential risks as a “distraction” that is “not grounded on science,” comments that conflict with warnings given to the United Nations by the heads of OpenAI and Anthropic.2 min read
SafetyAI and Automated Bots Play a Significant Role in Israel’s Election Campaign for the First TimeGuy Tytunovich, CEO of Tel Aviv–based cybersecurity firm CHEQ, says artificial intelligence and automated bots are already influencing Israel’s current election campaign.2 min read
SafetyYale neuroscientist Martin Jan Stránský warns of an AI-driven catastrophe within three yearsMartin Jan Stránský, a Yale professor and neuroscientist, says the rapid development of artificial intelligence driven by financial incentives makes some kind of catastrophe inevitable and predicts it could occur within three years.3 min read
SafetyAnthropic says Bay Area lab discovered a CRISPR-like enzyme system using its Claude AIAnthropic announced that its Bay Area wet lab, which uses the Claude AI to assist research, identified a previously unknown enzyme system in bacteriophage DNA with properties “reminiscent of CRISPR.”…3 min read
SafetyAI-driven biosecurity arms race: Radical Numerics and Eric Nguyen on defense and genome language modelsRadical Numerics, co-founded by Eric Nguyen and other AI×Bio researchers, is scaling genomic language models (GLMs) to tackle a broad set of biological problems — and warns that the same capabilities that accelerate biology also change the biosecurity landscape.4 min read
SafetyConcerns Rise as AI Systems Demonstrate Misuse and Prompt Calls for RegulationRecent incidents indicate AI agents have been used to exploit systems and obtain answers to tests, and companies report models behaving in ways that resemble unauthorized access.2 min read
SafetyThe company pledges to support independent, in-depth evaluationsAccording to the statement, the company commits to providing independent evaluators with broad access to models' training, evaluation and deployment processes so they can challenge assumptions,…1 min read
SafetyAI Hype vs. Reality: Marketing, Negligence and Overstated BreakthroughsRecent months have seen a string of high-profile AI claims and incidents—from Anthropic’s assertions about Claude Mythos and alleged mathematical breakthroughs to OpenAI’s controversial episodes and a viral resignation by an Anthropic engineer.4 min read
SafetyResearcher: Z.ai's ZCode Uploaded Developers' Local Workspaces by DefaultA developer who reverse-engineered Z.ai's local AI coding tool ZCode reports that the application silently snapshot and uploaded entire developer workspaces to Z.ai's cloud when users were logged in.2 min read
