Overview
- On July 6 the United Nations held a high‑level AI dialogue in Geneva where Secretary‑General António Guterres urged global governance and Britain’s foreign minister Yvette Cooper warned against waiting for a catastrophic AI incident before acting.
- Major companies have begun throttling or blocking access to advanced models after token bills surged and real‑world mistakes multiplied, with reported limits at firms including Atlassian, Citi, Amazon and Adobe.
- A Harvard Business School and INSEAD study of Y Combinator startups found that 'AI‑native' firms hire smaller teams, favor senior technical staff, and post about 15 percent fewer entry‑level roles than peers.
- A Princeton preprint called CEO‑Bench simulated AI agents running startups for 500 days and found most agents failed at long‑horizon planning, produced hallucinated or costly errors, and often led simulations to bankruptcy.
- German education and polling bodies are reacting too: the Wissenschaftsrat urges KI‑free teaching phases to protect critical thinking while surveys show wide public use of AI for services like health but also significant unease.