Researchers and executives from Anthropic, OpenAI, Google DeepMind and other leading AI labs issued a wave of stark warnings this week about the technology's catastrophic potential. Deep learning pioneer Yoshua Bengio backed an Anthropic researcher's resignation over safety fears, saying frontier lab scientists see risks months before the public does. He also cautioned that current AI systems already possess hacking and persuasion capabilities that could be weaponized against humans.
Fleet Device Management has been admitted to Project Glasswing, Anthropic's controlled-access program for its Claude Mythos Preview model. Fleet will use the early access to explore how frontier AI could help IT teams detect and remediate device vulnerabilities faster, aiming to shrink the delay between when threats appear and when organizations patch them.
Rune Kvist, a former Anthropic employee, and Rajiv Dattani, ex-COO of AI safety group METR, launched a startup called Artificial Intelligence Underwriting Company (AIUC) that builds a third-party audit and certification system for AI agents, modeled on the SOC 2 cybersecurity standard. The company, whose clients include Cursor, Lovable, Harvey, and ElevenLabs, just raised a $40 million Series A led by Ribbit Capital with First Harmonic, adding to a prior $15 million seed round, bringing total funding to $55 million.
President Trump used social media posts to reject growing concerns about AI expansion and data centers, dismissing them as a 'hoax' and 'scam' while crediting AI investment for America's tech lead. Separately, Microsoft released a draft code of conduct that would impose limits on future AI models, following calls from Anthropic and OpenAI for a slower pace of AI development. Meanwhile, Bank of America CEO Brian Moynihan warned that investment banking fees could drop more than 10% in the third quarter after a strong prior quarter.
Digital Realty CEO Andrew Power said warnings of a potential AI development slowdown from firms like Anthropic, OpenAI and xAI would not mean a halt to data center real estate demand. He argued that non-AI digital transformation and cloud computing growth remain strong drivers for his business, even as Digital Realty and Equinix shares fell on Monday amid AI slowdown fears.
OpenAI and Anthropic executives have begun publicly urging caution about AI's risks even as their companies pursue trillion-dollar valuations, prompting MIT Technology Review to examine what a real slowdown would look like and whether the industry's own calls for restraint can be trusted. The publication is also hosting a subscriber roundtable discussing the credibility of AI extinction fears, once considered fringe but now voiced by researchers within top labs.
OpenAI's charitable foundation, which holds a 26% stake in OpenAI worth up to $250 billion, has given $500,000 to nonprofit 1Day Sooner to acquire biological research datasets from companies that have gone bankrupt. The group's leader says such datasets could often be obtained for only tens of thousands of dollars, and the funding is meant to test whether this rescue model works at scale. This grant is part of the foundation's broader push to ramp up giving, including a $100 million donation earlier this year for hepatitis C treatment access.
A Warsaw-based user reports that on September 1, 2026, Google shut down a Gmail account they had held since 2011, citing spam without identifying any offending message. Despite five appeals, three registered letters, complaints to three public authorities and a visit to Google's Polish office, the company gave no explanation, while the user notes Google Poland is not party to any agreement covering the account.
Following a weekend of AI executives calling for slower development and stronger oversight, President Trump publicly rejected the idea of new guardrails, saying only a capable president is needed to control AI. Anthropic CEO Dario Amodei had proposed embedding independent evaluators and setting cross-government safety standards after two safety researchers resigned citing existential risk concerns, a position echoed by OpenAI's Sam Altman and Elon Musk.
NYU mathematician Tristan Buckmaster alleges OpenAI's announcement that its Astra model solved the century-old Navier-Stokes problem came suspiciously soon after he and Anthropic researcher Levent Alpöge developed their own solution using AI tools. Buckmaster says OpenAI scientists approached him to jointly announce the finding and allegedly suggested refusing could hurt his career, while declining to confirm whether their model had access to his research. OpenAI credited Astra with the proof and denied using customer data directly, though it admitted it could not rule out use of anonymized data to improve its models.
Investors are increasingly scrutinizing whether the massive data center buildout tied to Anthropic and OpenAI's growth can continue at its current pace, given uncertainty around future AI model progress. Companies like Oracle, GE Vernova, Caterpillar, Vertiv, Nebius, CoreWeave, Dell and Hewlett Packard Enterprise have all made major bets tied to sustained AI infrastructure demand.
Fast Company's latest coverage examines how major companies including Walmart, Meta, Anthropic, and Adidas are ramping up the pace at which they adopt and deploy AI tools across their operations. The piece follows an earlier editor's note describing executives as struggling to keep pace with AI-driven change, based on a CEO survey the publication conducted.
GoKawiil