Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
A study found that AI answered more than 80% of U.S. college entrance exam questions in an environment similar to .. The finding comes amid a series of test irregularities involving AI smart glasses.
Read the original →Debt collection agencies are using AI to calculate from whom there is something to recover, and the Dutch Data Protection Authority is sounding an alarm on the practice. Computer systems reportedly determine, based on place of residence, credit score and payment history, which people to contact first.
Read the original →CPAs must understand AI's limits, test its outputs, and keep professional skepticism firmly in human hands.
Read the original →This is not the first time Anthropic has issued a watermark. The latest iteration will apply to Mythos 5, Opus 4.7, Opus 4.6, and Sonnet 4.6.
Read the original →The growing use of generative AI in political and social movements continued to pose new challenges for fact-checkers in October. Examples range from fabricated protest visuals to gender-swapped videos.
Read the original →A new infostealer uses AI to compile dossiers from stolen credentials to receipts.
Read the original →The main points of ISD Policy 625 on the Responsible Use of AI are that it is permitted but regulated, used as a supplement and not a replacement, and that teachers retain control over classroom use.
Read the original →