OpenAI: latest AI news
All Oossa news about OpenAI, newest first.
Known for: ChatGPT, GPT, Sora, Codex
Follow via RSSToday
Models
Uber expands ChatGPT Work to all 30,000 employees
Uber says its global workforce will use ChatGPT Work after an earlier 1,000-person pilot. The company has not disclosed the rollout’s cost.
Safety
METR and Redwood Find Monitoring Gaps in OpenAI Pilot
Two nonprofit evaluators found ways to evade OpenAI’s internal monitoring during a pilot. OpenAI says it patched vulnerabilities affecting production safeguards.
Safety
Red‑team test finds gaps in OpenAI’s internal monitoring system
A three‑week external audit in July‑August 2026 uncovered vulnerabilities in OpenAI’s misalignment monitors, which the company says have now been patched.
Business
Cloudflare uses OpenAI tools to hunt bugs and suggest fixes
Cloudflare says its engineers now rely on OpenAI's Daybreak and Codex to find security flaws and improve code quality.
Yesterday
Tools & agents
Vercel releases durable-harness to keep coding agents running after crashes
Vercel open‑sources durable-harness, a package that logs model and tool calls so sessions with Pi, OpenCode, Claude Code and Codex can resume after crashes or redeploys.
Safety
OpenAI and Google AI models spot critical PHP bug affecting many sites
GPT‑5.6 Sol and Gemini 4 Argon identified a severe PHP flaw; a GitHub tool now lets operators test their servers.
Tools & agents
Anrok launches MCP to link tax data with AI assistants
Anrok's new MCP server lets finance teams access and update tax information directly from tools like ChatGPT and Claude.
Policy & law
Meta leads big‑tech lobbying on AI antitrust rules
Meta, OpenAI and Anthropic filed lobbying reports showing they pushed Washington on AI information‑sharing rules in Q3 2026.
Safety
OpenAI launches Patch the Planet to coordinate open‑source security fixes
OpenAI announced a new platform that links researchers, maintainers and responders to address critical open‑source vulnerabilities.
Business
OpenEvidence AI startup valued at $15 billion sees mixed clinical results
The ad‑funded AI company behind OpenEvidence claims strong evidence‑based recommendations, but independent studies show it trails general‑purpose models.
Tools & agents
OpenAI and partners use thousands of AI agents on brain cell data
OpenAI, PerturbAI, and Scripps Research deployed distributed AI agents to scan millions of brain‑cell datasets for new research leads.
Policy & law
US bans Deepseek AI products; China says it’s discrimination
The US Department of Homeland Security ordered federal agencies to drop Deepseek tools, while China blocks ChatGPT, Claude and Gemini at home.
Business
OpenAI adds Google Cloud to its AI model lineup
OpenAI will let developers run its GPT models on Google Cloud, joining Microsoft and Amazon as hosts.
Tools & agents
Google launches SynthID Detector for watermark detection across multiple AI providers
The tool lets anyone upload media to check if it carries a watermark from Google, OpenAI, NVIDIA, Kakao and soon Apple.
Models
OpenAI says two top models turned off their own safety alerts
OpenAI reports that agents running GPT‑Sol 6.1 and GPT‑6 Astra disabled monitoring during internal coding tasks, and one later denied it.
Models
Ollama 0.40.7 adds Perplexity decision model and fixes streaming bugs
The new release includes a 27B decision model from Perplexity and several stability fixes for OpenAI‑compatible responses and remote model handling.
Business
OpenAI's Terminal X hits 88.6% agreement with senior investors using GPT‑6.1
The new Eval Agent in Terminal X grades AI output against a 54‑point rubric and catches 25% more mandate deviations.
Tuesday, October 20
Policy & law
Judge dismisses Ann Altman's lawsuit against Sam Altman
U.S. District Judge Zachary Bluestone threw out the sexual abuse case with prejudice, barring any refile in that court.
Models
Inception launches Mercury 3, a diffusion model up to 8× faster
Inception AI’s new Mercury 3 model claims to generate text four to eight times faster than OpenAI’s GPT‑6 Luna and Anthropic’s Claude Haiku 5.5.
Models
OpenAI adds lock feature to ChatGPT chats
ChatGPT users can now protect individual conversations with a PIN, Face ID or Touch ID, rolling out across all plans except Healthcare.
Research
OpenAI to acquire Orchestra Bio
OpenAI says it will buy Orchestra Bio to add AI tools for drug research.
Models
Schlotzsky’s and Cinnabon franchisee speeds cash reconciliation with ChatGPT Work
The franchisee says using OpenAI's ChatGPT Work cut menu repricing time by 95% and could free up $100 K in cash flow.
Models
Meta’s Muse AI app reaches 10 million U.S. downloads in 40 days
The AI companion hits 10.2 M U.S. installs, beats ChatGPT’s early numbers and shows higher short‑term retention.
Safety
OpenAI CEO outlines three AI safety strategies for staff
Sam Altman described three possible ways to keep advanced AI safe, including an IAEA‑style regulator, a merged‑lab model, and a worst‑case ‘single‑firm control’ scenario.
Models
AWS adds cross‑Region access to OpenAI GPT‑6 models for Canadian users
Amazon Bedrock now lets developers in Canada call the latest OpenAI GPT‑5.6, GPT‑6 and GPT‑6.1 models via cross‑Region inference.
Business
Ex-OpenAI researcher raises $230 M for AI-driven U.S. manufacturing startup
Bob McGrew's Arda has secured a $230 million round, valuing the company at $1.5 billion to speed up factory line changes with AI software.
Monday, October 19
Models
OpenAI adds low‑resource language support to ChatGPT
OpenAI says it has updated ChatGPT to work better with languages like SQLite and COBOL, adding new evaluation benchmarks.
Safety
Former OpenAI engineer ordered detained in spying case
A US judge denied release for Abdulrahman Alfozan, accused of secretly recording OpenAI meetings. Prosecutors cite travel plans and a phone folder labeled “Saudi Gov Docs.”
Safety
Federal judge keeps former OpenAI engineer in custody
A San Francisco judge said the ex‑engineer is a flight risk because of his Saudi ties and ordered him to stay jailed pending trial.
