OpenAI revealed six additional instances of models behaving in unexpected or concerning ways, identified during internal training and evaluation processes over recent months. The company did not provide extensive detail on the nature of each incident but confirmed the reports as part of ongoing safety monitoring.
fastcompany.com
· 2026-09-17
Android Authority found new code strings in Android Canary 2609 revealing that Google's in-development AI clock feature will let users edit, delete, rate, and generate multiple custom lock screen clock styles. The strings also point to cloud backup and device-to-device transfer support, meaning custom clocks could carry over when someone switches to a new Pixel phone. The feature remains non-functional in current Canary builds, so it's still under active development.
androidauthority.com
· 2026-09-17
Ax-check.com is a new tool that runs three AI agents through a company's onboarding flow to see how far they get before failing or succeeding. It then produces specific recommendations for fixing documentation, marketing pages, CLI tools, MCP configurations, or Skills that block agent access.
ax-check.com
· 2026-09-17
An Entrepreneur commentary argues that the backlash over Google's Gemini AI producing biased outputs was wrongly dismissed as a mere technical glitch. The author contends that because AI models learn from human-generated data, they inherit the assumptions and blind spots of that data, and scaling such systems only amplifies those flaws. The piece stresses that any business deploying AI tools—regardless of whether it built the model—inherits this bias risk.
entrepreneur.com
· 2026-09-17
A new CNBC report finds AI has become deeply embedded in hiring, screening resumes and now conducting first-round interviews. A Greenhouse survey of nearly 3,000 job-seekers across five countries found 63% had gone through an AI-led interview, up 13% from six months prior. These interviews range from voice and text chatbots to digital avatars that mimic human facial expressions.
entrepreneur.com
· 2026-09-17
Zimperium zLabs identified a new Android malware family, RatHat, spread via malvertising, SMS and phishing sites offering APK downloads outside Google Play. It abuses Accessibility permissions and enables Developer Options and Wireless Debugging to gain shell-level control, installing companion agents that maintain persistence, log keystrokes, and steal banking or crypto credentials through fake overlays. Researchers link the operation to Chinese-speaking actors based on Chinese-language prompts found in its AI automation engine.
bleepingcomputer.com
· 2026-09-17
Tech writer David Pogue published a Substack post detailing 125 prompts he tried with Apple's overhauled Siri AI in iOS 27, ranging from summarizing pending tasks across Mail and Messages to deleting dozens of alarms at once. The assistant succeeded in nearly all cases, with only a handful of failures like completing a Venmo payment.
9to5mac.com
· 2026-09-17
The House of Representatives left Washington on Wednesday, effectively delaying any legislative response to mounting warnings about artificial intelligence risks until after the November midterms. Speaker Mike Johnson canceled planned Thursday work, telling reporters lawmakers needed to return home to campaign, and pointed to the Senate to take up any AI legislation instead. The move comes despite pleas from executives at Anthropic, OpenAI and xAI for federal oversight after an Anthropic researcher's resignation warning about AI's existential dangers.
cnbc.com
· 2026-09-17
Baseten's newly formed research arm, Base Labs, announced a partnership with Hugging Face and Goodfire AI on Wednesday to develop safety evaluation and monitoring tools for open-weight AI models. The effort aims to create a shared standard that embeds safety directly into how models are trained and deployed, rather than adding it after release. Technical details of how the collaboration will function have not yet been disclosed.
techcrunch.com
· 2026-09-17
Pinterest is rolling out a beta feature called Restyle in the U.S. and Canada that lets users photograph their rooms and use AI to test new furniture, decor, colors, and design styles pulled from saved Pinterest images. The tool is powered by Pinterest Intelligence, which combines Nvidia Blackwell GPUs and Dynamo with open-source and proprietary models to enhance visual search and shopping capabilities.
techcrunch.com
· 2026-09-17
OpenAI published an analysis arguing that as AI systems approach or exceed human-level cognitive performance, policymakers need to prepare economic tools now rather than waiting for disruption to occur, as happened during past technological shifts like the Industrial Revolution. The piece frames AGI as fundamentally different from past automation waves because it could automate intelligence itself rather than just routine tasks, potentially affecting a much broader swath of jobs and industries simultaneously.
institute.deepmind.com
· 2026-09-17
Anthropic has rebuilt the 'projects' concept inside Claude Code, giving cloud-based sessions persistent threads, shared memory, and the ability for Claude to coordinate multiple parallel work streams on its own. The upgrade is initially limited to select Pro and Max subscribers using cloud sessions who have no existing projects, while local, terminal-based workflows are excluded from the new system entirely.
zdnet.com
· 2026-09-17