News — 2026-05-17

A tech writer discovered that adding a skeptical instruction to ChatGPT prompts forces the model to flag uncertain answers, cutting down on fabricated details. By telling the AI to act as a hostile auditor, the author saw more cautious replies across travel planning, appliance troubleshooting, and product advice, though the model still isn’t immune to errors. Read more

OpenAI announced a partnership with Malta that will give every resident and citizen a free year of ChatGPT Plus, provided they complete an AI basics course offered by the University of Malta. The program, described as the first of its scale, requires participants to hold an EU eID and will be administered by the Malta Digital Innovation Authority. OpenAI also revealed it is pausing its planned Stargate data center in the United Kingdom, citing high energy costs and regulatory hurdles. Read more

OpenAI announced that co‑founder and president Greg Brockman will lead the company’s product strategy on an interim basis while CEO Fidji Simo remains on medical leave. Brockman outlined a plan to merge ChatGPT with the Codex programming tool into a single experience, signaling a shift toward a unified AI “super app” after a recent internal “code red” and the suspension of several side projects. Read more

OpenAI unveiled a new suite of personal finance features on Friday, making them available in preview to ChatGPT Pro subscribers in the United States. The tools let users link bank accounts, credit cards and investment platforms through a partnership with Plaid, then ask the chatbot to analyze spending, track subscriptions and model future financial goals. The rollout follows OpenAI’s acquisition of the Hiro finance‑tech team earlier this year and leverages the latest GPT‑5.5 model, which OpenAI says improves contextual reasoning for money‑related queries. Read more

A growing number of people are feeding AI tools with old messages, photos and social‑media data to create digital versions of former lovers. The practice, which began on the open‑source platform Colleague.skill, raises questions about consent, privacy and the impact on grief, while some users claim the bots helped them process breakups. Read more

OpenAI is reportedly consulting lawyers after Apple’s rollout of ChatGPT within Siri failed to meet the AI firm’s expectations. Insiders say the partnership, once billed as a multi‑billion‑dollar opportunity, suffered from design choices that made the service hard to use and poorly promoted. OpenAI now considers legal options, citing a lack of transparency and what it calls an “honest effort” from Apple. Read more

The Commodity Futures Trading Commission announced a crackdown on U.S. traders slipping into offshore crypto prediction platforms like Polymarket. Chair Michael Selig said the agency is expanding staff and deploying artificial‑intelligence tools to spot suspicious betting patterns, while also using blockchain‑analysis services such as Chainalysis. The move signals a new enforcement push targeting insider trading and market manipulation in the rapidly growing prediction‑market space. Read more

OpenAI announced a sweeping reorganization that folds its flagship products—ChatGPT, Codex and the developer API—into a single product organization led by co‑founder Greg Brockman. The move, detailed in an internal memo, aims to eliminate redundant engineering effort, focus resources on a unified "agentic" platform and streamline the company’s narrative ahead of a planned IPO. The restructuring follows a series of shutdowns dubbed "side quests," a December "code red" alert and ongoing legal battles that have pressured the firm to tighten its compute budget and sharpen its market position. Read more

A Bitcoin investor who thought his five coins were gone forever reclaimed almost $400,000 after Anthropic’s Claude AI identified a forgotten wallet backup on an old college computer. The user, who lost the password more than ten years ago, uploaded years of scattered files to the chatbot. Claude pinpointed an earlier backup file and, combined with a rediscovered mnemonic phrase, unlocked the wallet without breaking Bitcoin’s encryption. The episode underscores how generative AI can serve as a practical tool for navigating tangled digital archives. Read more

A San Francisco federal judge refused to give final approval to Anthropic's $1.5 billion settlement with authors who accuse the AI firm of training its Claude models on pirated books. Judge Araceli Martínez‑Olguín asked for a detailed breakdown of counsel fees, lead‑plaintiff payments and other expenses before signing off, leaving the largest proposed copyright settlement in U.S. history in limbo. Read more

Former OpenAI chief technology officer Mira Murati unveiled a new class of AI at Thinking Machines Lab that prioritizes human collaboration. The startup demonstrated “interaction models” that process live audio‑visual input, understand pauses and tone, and respond in real time, a stark departure from conventional voice assistants that simply transcribe speech. While the prototypes remain unreleased, Murati says the technology is meant to keep people in the loop as AI grows more capable. The move underscores a growing divide in the industry between firms racing toward autonomous superintelligence and those betting on human‑in‑the‑loop designs. Read more

Cerebras Systems closed its first day on the Nasdaq at $311.07, up 68% from its $185 IPO price, giving the wafer-scale chip company a market capitalization of approximately $95 billion. The offering raised $5.55 billion, making it the largest US tech IPO since Snowflake's $3.8 billion debut in 2020. This milestone validates investor appetite for pure-play AI hardware and sets the stage for potential IPOs from SpaceX, OpenAI, and Anthropic, worth $3 trillion combined. Read more

A recent social media post from Menlo Ventures partner Deedy Das highlights the stark contrast between those who have made fortunes in the AI industry and those who are struggling. Das estimates that around 10,000 people, including founders and employees at companies like OpenAI and Nvidia, have achieved 'retirement wealth' of over $20 million, while others are facing layoffs and uncertainty about their career paths. Read more

Finals season can be overwhelming, but AI tools are here to help. Students can use AI to build a study plan, manage their time, and stay on track. From estimating assignment completion times to providing reminders and emotional support, AI tools like Copilot, Gemini, and Abby can help students succeed. Read more