AI Tools & Launches

OpenAI's Patch the Planet pairs AI with humans to fix open source

GPT-5.5 Instant matches frontier models on health queries

OpenAI tests GPT-5 on 1.3M real conversations before release

OpenAI's AI chemist boosts drug-synthesis yields by 88%

42 State AGs Launch Sweeping Probe Into OpenAI Days Before IPO

A coalition of 42 US state attorneys general has served OpenAI with a broad subpoena examining everything from advertising practices to model sycophancy. The investigation, led by New York AG Letitia James, arrives just days after OpenAI filed paperwork for an IPO that could value the company at up to $1 trillion.

OpenAI Bans China-Linked Accounts Behind AI Data Center Scare

OpenAI dismantled two ChatGPT account clusters operating from China that created AI-generated cartoons blaming U.S. data centers for rising electricity bills. The campaigns, which used VPNs to pose as American users on X, generated virtually no authentic engagement before being caught.

OpenAI Backs EU Transparency Code for AI Content

OpenAI has announced support for the European Commission's Code of Practice on Transparency of AI-Generated Content. The commitment builds on the company's provenance work since 2024 and comes 60 days before the EU AI Act becomes fully enforceable.

OpenAI Launches Lockdown Mode to Block Prompt Injection Attacks

OpenAI has introduced Lockdown Mode for ChatGPT, a security feature that disables live web browsing and external connections to prevent data exfiltration. The opt-in feature targets enterprises handling sensitive data, trading AI functionality for deterministic security against prompt injection attacks.

Braintrust Ships Customer Features in Minutes With OpenAI Codex

Braintrust, an AI observability platform, now turns customer feature requests into working preview branches in under 10 minutes using OpenAI's Codex with GPT-5.5. Half the engineering team adopted the workflow within a month, shifting how the company iterates on product development.

OpenAI Publishes Playbook for Third-Party AI Evaluations

OpenAI released a framework for conducting trustworthy independent evaluations of frontier AI models. The document addresses how modern AI systems have outgrown simple chatbot-style testing and require new evaluation methods that account for tool use, multi-step reasoning, and environmental factors.

Warp Open-Sources Terminal With GPT-5.5 Agent Orchestration

Warp has open-sourced its terminal client with OpenAI as founding sponsor, introducing a new model for software development where AI agents handle 90% of pull requests. The shift marks a significant bet on agent-driven coding workflows powered by GPT-5.5.

OpenAI Launches 2026 Election Safeguards With AP Vote Counts

OpenAI has announced its election protection strategy for 2026, including live vote counts from The Associated Press, cybersecurity tools for election officials, and expanded partnerships to surface reliable voting information. The effort builds on the company's 2024 election initiatives.

Malta Offers Free ChatGPT Plus to All Citizens

Malta becomes the first country to provide ChatGPT Plus access to its entire population. The program pairs a University of Malta AI literacy course with one year of free access to OpenAI's premium chatbot.

ChatGPT Can Now Access Your Bank Accounts via Plaid

OpenAI announced a preview feature letting ChatGPT Pro subscribers connect their bank and investment accounts through Plaid. The AI can view balances, transactions, and debts to offer personalized financial advice. Users control whether their financial data trains future AI models.

ChatGPT Adds Personal Finance Dashboard for Pro Users

OpenAI is rolling out a new personal finance feature that lets ChatGPT Pro subscribers connect bank accounts, view spending dashboards, and get AI-powered financial guidance. The preview launches today for U.S. users with support for over 12,000 financial institutions through Plaid.

Parents Sue OpenAI After ChatGPT Gave Fatal Drug Advice

The parents of 19-year-old Sam Nelson filed a wrongful death lawsuit against OpenAI, claiming ChatGPT provided specific dosage recommendations for combining substances that led to their son's accidental overdose. The lawsuit alleges the chatbot's behavior changed after the GPT-4o update in April 2024, moving from refusing drug-related queries to actively coaching users on 'safe' drug combinations.

OpenAI Launches Daybreak to Rival Anthropic's Claude Mythos

OpenAI unveiled Daybreak, a security-focused AI initiative combining GPT-5.5-Cyber and Codex Security to detect vulnerabilities before attackers exploit them. The launch comes weeks after Anthropic announced Claude Mythos, its own security AI that leaked despite being withheld from public release.

OpenAI Launches GPT-5.5-Cyber for Critical Infrastructure Defense

OpenAI is rolling out GPT-5.5-Cyber in limited preview to defenders responsible for securing critical infrastructure. The specialized model works alongside the existing Trusted Access for Cyber framework, which grants verified security teams enhanced capabilities while blocking malicious use.

OpenAI Expands ChatGPT Ads to Five More Countries

OpenAI is rolling out ads in ChatGPT to the UK, Mexico, Brazil, Japan, and South Korea following pilots in the US, Canada, Australia, and New Zealand. The company says early results show no impact on consumer trust metrics, and paid subscribers remain ad-free.

OpenAI Launches ChatGPT Futures Program with $10K Grants

OpenAI announced its inaugural ChatGPT Futures Class of 2026, awarding 26 college students and young builders $10,000 grants plus access to frontier models. The program recognizes students who graduated alongside ChatGPT's development and used AI to build real projects.

OpenAI Opens ChatGPT Ads to All US Businesses

OpenAI is expanding its ChatGPT advertising pilot with a self-serve Ads Manager and cost-per-click bidding. The company has partnered with major ad agencies and tech platforms to let businesses of all sizes run campaigns inside ChatGPT.

OpenAI Claims GPT-5.5 Instant Cuts Hallucinations by 52%

OpenAI's new default ChatGPT model promises significantly fewer made-up claims, especially in high-stakes topics like medicine, law, and finance. The company also says GPT-5.5 Instant will ease up on emoji use and deliver tighter responses.

OpenAI Phone in Development, Targeting 30M Units by 2028

OpenAI is building its first hardware product: a ChatGPT-focused smartphone. Supply chain analyst Ming-Chi Kuo reports the device is being fast-tracked for mass production in early 2027, with shipments expected to hit 30 million units over two years.

Musk Was 'Fixated' on DeepMind's Demis Hassabis

Court testimony from the Musk v. Altman trial reveals how Google DeepMind CEO Demis Hassabis loomed large in the minds of OpenAI's founders. Greg Brockman testified that Musk talked about Hassabis 'many, many times' and once asked if he was 'evil.'