The ai law & regulation stories our desk judged worth your time, written so anyone can follow them.
Anthropic found Claude could write training data that passed a safety screen yet taught another AI to blackmail and sabotage code, in tests only.
Read more: Anthropic Shows Claude Can Sneak Bad Habits Into Another AIOPay called off its planned New York Stock Exchange debut over a regulator's inquiry, but says its app and payments still work normally.
Read more: OPay pulls its New York stock market debut over a regulator's inquiryThe White House declined to hand over key papers on its voluntary AI testing plan, even after agreeing in September to release documents. Here is why that matters.
Read more: White House withholds papers on its voluntary AI safety testing planFreecash, an app that pays small sums for playing games, briefly hit the App Store's top five, but users say the pay is thin and spending traps are easy to fall into.
Read more: Apps That Pay You to Play Games Are Hot, but Most Users Wash OutFilings with the SEC show 47 companies now mention humanoid robots, and the article says past buzzword surges often came before market bubbles.
Read more: 47 companies now mention humanoid robots in SEC filingsServiceNow says buying Israeli startup Sweep lets big companies switch customer software in 30 to 60 days. That could change how hard it is to leave a vendor.
Read more: ServiceNow buys Sweep to cut customer software switches to weeksCriminals now use AI-made video, cloned voices and fake images to pose as celebrities, bosses and relatives. A familiar face or voice no longer proves who is real.
Read more: Fake videos and cloned voices are now used to steal moneyAn analysis says candidates who back a data center pause could win up to 15 of 36 governor races on Tuesday, which could shape where new ones get built.
Read more: Data center pause is now a campaign issue in up to 15 governor racesA family says ChatGPT helped their daughter taper off epilepsy medicine before she died. OpenAI calls it tragic. The case is a warning about asking chatbots for medical advice.
Read more: Parents Sue OpenAI, Say ChatGPT Helped Daughter Stop Epilepsy MedicineA Vals AI study of 25 chatbots found Chinese models lean toward Beijing less than expected, and the language you ask in can change the answer.
Read more: Chinese AI chatbots lean toward Beijing's views, but less than expectedIndia's EPFO plans to pay out inactive provident fund accounts holding up to ₹1,000 automatically, so small leftover balances need no claim.
Read more: India's retirement fund to auto-close tiny inactive accountsNew York paused new data center building for a year. Residents and lawmakers say it falls short, and many other states may soon face the same pressure.
Read more: New York paused new data centers. Locals say that is not enoughCheap AI voice tools are being sold to criminals, and a call that sounds like your family may not be them. Here is what to know and what to watch for.
Read more: Scammers can now copy a loved one's voice for as little as $8AskAnyModel AI Pro is a lifetime plan that puts answers from several chatbots in one window. It is on sale for $29.97, but check the credit limits first.
Read more: One $30 tool shows answers from 50 AI chatbots side by sideA federal judge told OpenAI to keep all ChatGPT chats, even deleted ones, so anything you type may now be stored while the court case continues.
Read more: Court orders OpenAI to keep all ChatGPT chats, even deleted onesCNET's review finds PrivadoVPN among the cheapest VPNs around and easy to use, but it has no outside security audits and slows internet speeds a lot.
Read more: PrivadoVPN is very cheap and easy to use, but has no outside auditsTuesday's election may show whether the White House push for AI and data centers is costing the president support among his own voters.
Read more: Tuesday's election tests whether Trump's AI push is costing him votesA federal judge refused to dismiss the NAACP's suit saying xAI ran gas turbines at its Memphis-area data center without required air permits.
Read more: Judge lets NAACP lawsuit over xAI's Memphis data center pollution proceedGeneral Motors will pay $220 million to settle a case over selling driving data to insurers, a reminder that how you drive can follow you to your bill.
Read more: GM will pay $220 million over selling drivers' data to insurersAnthropic is adding a hidden text watermark to four more Claude models, but on two of them one settings change can switch it off.
Read more: Anthropic marks Claude's text with a hidden pattern, but two models have a gapOne Apple IT writer argues that heavy AI users could save money by running models on a Mac instead of paying cloud bills that grow with every task.
Read more: Running AI on your own Mac may cost less than paying per useAlphabet, Amazon and Intel fell on worries about AI spending, while Ford and GM rose. A possible December rate hike adds to the pressure on your investments.
Read more: Big tech stocks slide as AI spending worries grow, but Ford and GM riseExperts say the old tricks for spotting AI fakes, like counting fingers, no longer work. Checking the source and your own reaction is now the better defense.
Read more: Spotting fake AI images by their flaws no longer works, experts sayHuge AI data centers can push up electric bills for people who live miles away, and some states are now making the companies pay more of the cost.
Read more: AI data centers can raise your electric bill even if none is nearby
