A for AI

News Archive

Pick a date to read that day's stories.

July 2026

Days with a blue dot have stories.

DISPATCH
Monday, July 27, 2026
  1. 01 of 10NEWS

    Nvidia and Microsoft build an AI safety club that snubs the big three

    Nvidia, Microsoft and 35 other firms formed an open AI security alliance so defenders can run their own protective tools instead of relying on OpenAI, Google or Anthropic. It follows worries about AI systems being used to attack companies automatically.

    OUR TAKEEven robots need friends who are not the popular kids.

    Read more →
  2. 02 of 10NEWS

    Europe's AI rulebook gets its first big rewrite

    The EU's Digital Omnibus on AI took legal effect on 27 July 2026, easing some compliance burdens on companies while keeping bans on manipulative and deepfake nudifier apps due by December. It is the first change to the bloc's AI law since 2024.

    OUR TAKEBrussels simplifies the rules, which itself needs its own rulebook.

    Read more →
  3. 03 of 10NEWS

    Brazilian bank's AI chat assistant now serves 300,000 customers

    Itau, one of Brazil's largest banks, widened access to its GenAI assistant iAi to 300,000 clients on 27 July 2026, letting people ask everyday questions about their accounts in plain language. It plans to reach 3 million users by September.

    OUR TAKENice to see a bank chatbot that might actually answer your question.

    Read more →
  4. 04 of 10NEWS

    Brazilian publisher bets big on homegrown school AI

    Textbook publisher FTD announced on 27 July 2026 it will spend 100 million reais a year building AI tools for schools without using systems from the large American tech firms. The Marista Group company wants education software built and controlled locally.

    OUR TAKEA rare case of someone reading the terms and conditions first.

    Read more →
  5. 05 of 10NEWS

    Chinese humanoid robots start real factory shifts

    Construction giant Zoomlion put dozens of its own humanoid robots to work on real production lines in Changsha, testing them on nearly 20 tasks including assembly and quality checks. It is one of several Chinese firms moving robots from demos to actual factory floors.

    OUR TAKEThe robots clocked in before most of us finished our coffee.

    Read more →
  6. 06 of 10NEWS

    South Korean hospital's AI beats humans on medical exams

    Samsung Medical Center's in-house AI tool ZEO Med 2 scored 97.7 on Korea's medical licensing exam and 94.8 on the US equivalent, the highest of any tested system. It is freely available on Hugging Face, a website where anyone can download AI models.

    OUR TAKEIt aced the exam, though bedside manner remains untested.

    Read more →
  7. 07 of 10TOOLS

    Seychelles fintech builds an AI operating system for African traders

    Fusepay, based in Seychelles, launched Fuse360, software built around AI that helps wholesalers and distributors across Africa manage stock, orders and payments in one place. It marks the payments company's move into wider business software.

    OUR TAKESmall island, big ambitions, and hopefully fewer spreadsheets.

    Read more →
  8. 08 of 10NEWS

    Brazil lets farmers borrow money using AI-tracked cows

    In a first for Brazil, a bank issued a rural loan backed by ten dairy cows fitted with collars that use AI to track their health and location, reducing the lender's risk. The digital record stops the same cow being pledged twice.

    OUR TAKEThe cows now have better data protection than most humans.

    Read more →
  9. 09 of 10NEWS

    Musicians widen their legal fight against AI song generators

    Thousands of independent musicians have joined US class action lawsuits against AI music platforms Suno and Udio, claiming their songs were used to train the tools without permission or payment. A musicians' union also filed an amended suit against major labels over related deals.

    OUR TAKETurns out robots writing hit songs still need human hits first.

    Read more →
  10. 10 of 10NEWS

    Fake AI version of Bolsonaro stirs up Brazil's election

    An AI-generated clone of former president Jair Bolsonaro's voice and likeness appeared without his authorisation during Brazil's election campaign, and the country's top electoral court is now deciding if it broke the law. His lawyer says he never approved it.

    OUR TAKEEven a fake Bolsonaro manages to cause a real headache.

    Read more →