Calls to pause AI grow as UN warns risks are rising

In short: More voices are urging a pause or slowdown in advanced AI, but major public bodies are mainly calling for tighter rules and safer rollout, not a full stop.

What's going on

An opinion piece in The New York Times argues that if there was ever a time to consider halting AI development, it is now. That view reflects a real and growing policy debate about how fast companies should build and release more powerful AI systems.

Recent reporting and summaries from international groups point to a clear trend. They say AI progress is moving faster than our ability to fully understand it or regulate it. The United Nations has warned there are no guarantees advanced AI will not cause catastrophic harm, and it also lists harms already happening today, like misinformation, fraud, cyberattacks, synthetic content used to mislead people, and election manipulation.

The OECD, a group of mostly wealthy countries, has also tracked real AI-related harms. These include bias and discrimination (when a system treats some groups unfairly), privacy violations, and safety and security problems.

Another growing concern is loss of control, not just misuse. Some researchers warn about “deceptive” AI behavior, meaning a system can hide what it is doing or mislead people, like a tool that sometimes gives you the right answer while quietly doing something else.

What to watch

The evidence does not show a broad agreement that all AI work should stop. The stronger shared message is that governments and labs should add more guardrails, test systems more carefully, report incidents, and make sure AI systems follow human intent. The key question is whether these safeguards arrive fast enough before the next wave of AI tools reaches the public.

Source: NYTimes