Fast AI Growth Brings Calls for Safety Rules

Artificial intelligence, or AI, is improving quickly. A major recent change is the rise of AI agents: programs that can use computer tools, browse the internet, and carry out tasks instead of only answering questions in a chat.

Supporters say these tools could help people with programming and other work. But some experts worry that agents may also create risks. They can search for mistakes, called bugs, in computer software much faster than people can. This could help teams fix their own systems, but criminals might try to use the same ability to attack systems.

OpenAI, Anthropic, Google, Meta, and companies in China are spending heavily to build more powerful AI. The competition has made some countries reluctant to slow their own AI work.

The source reports that OpenAI released a model called GPT-6 Astra last week. Its president said it may be an early step toward artificial general intelligence, or AGI. Some AI leaders, including Anthropic's leader on September 12, have called for slower development, monitoring, and protections against harm. Other leaders agreed, while U.S. President Donald Trump said a slowdown was not needed.

The European Union has made a major effort to create AI rules, but the source says its laws are not yet complete. For now, there is no single global group that can make and enforce AI safety rules for every company.

AI can bring useful new tools as well as serious challenges. Experts continue to debate how quickly it should grow and what safeguards are needed.