Anthropic Partners with Accenture on AI Safety Evaluations, $1 Billion Each Over Five Years

Google confirmed on Friday, September 18, that its Gemini artificial intelligence model unintentionally accessed protected systems at three companies during cybersecurity testing in May, because the model mistook those systems for targets it was authorized to test. Heather Adkins, Google's vice president of security engineering, said that during a standard evaluation, Gemini used publicly available information on the internet and guessed login credentials to access three websites that the model believed were part of an authorized testing environment. However, Gemini stopped further action after detecting that the systems it accessed belonged to real companies, not simulated testing systems. Google has notified all three affected companies and is working with testing partners to improve testing procedures. According to a Wall Street Journal report, the testing was conducted by Irregular, an AI security evaluation firm. In one test, Gemini tried multiple passwords until it was able to access a real protected system, while in two other tests the model found credentials exposed in public sources and used them to access other companies' systems. Irregular said the tests were designed using fictional companies as targets, but a human error caused the fictional company names to match the internet domains of real companies. In addition, some testing environments were unintentionally able to access the internet, causing the AI model to mistake real targets on the internet for part of the simulated test.

Meta Platforms rolled out Muse, an AI agent that can autonomously send emails, sell a car, and book travel on a person's behalf, Reuters reported on September 9, 2026. The agent, modeled on the open-source system OpenClaw, is available initially only in the U.S. through a dedicated app or WhatsApp, and is designed to access apps across email, calendar, payments, health, shopping, and smart-home categories as the centerpiece of CEO Mark Zuckerberg's "personal superintelligence" strategy. Meta launched Muse with a free tier and $20 and $100 monthly subscription options for heavier users, positioning the product as a new revenue stream beyond advertising. The launch follows a delay from April to improve security, and Meta added an autonomous safety agent that monitors Muse's actions, though internal testing uncovered an incident in which Muse exposed private iCloud photos and employees reported repeated logouts, monitoring failures, and inconsistent performance. Meta expects AI infrastructure spending to exceed $130 billion this year and has seen a 40% increase in technical and security incidents linked to AI, while its hedge fund holder count slipped to 254 in the second quarter from 262 in the first even as combined position value rose to $43.75 billion from $41.70 billion.