EU lawmakers call for summit to control 'very powerful' AI

EU lawmakers urged world leaders on Monday to hold a summit to find ways to control the development of advanced artificial intelligence (AI) systems such as ChatGPT, saying they were developing faster than expected.

The 12 MEPs, all working on EU legislation on the technology, called on US President Joe Biden and European Commission President Ursula von der Leyen to convene the meeting, and said AI firms should be more responsible.

The statement came weeks after Twitter owner Elon Musk and more than 1,000 technology figures demanded a six-month pause in the development of systems more powerful than Microsoft-backed MSFT.O OpenAI's latest iteration of ChatGPT, which can mimic humans and create text and images based on prompts.

That open letter, published in March by the Future of Life Institute (FLI), had warned that AI could spread misinformation at an unprecedented rate, and that machines could "outnumber, outsmart, obsolete and replace" humans, if left unchecked.

Response to the letter

The MEPS said they disagreed with some of the FLI message's "more alarmist statements."

"We are nevertheless in agreement with the letter's core message: with the rapid evolution of powerful AI, we see the need for significant political action," they added.

The letter urged democratic and "non-democratic" countries to reflect on potential systems of governance, and to exercise restraint in their pursuit of very powerful AI.

A spokesperson for von der Leyen did not immediately respond to a request for comment.

Responding to the lawmakers’ accusation of alarmism, FLI President Max Tegmark said: “They can think whatever they want as long as they do the right thing.”

Tegmark told Reuters he was “delighted” by the reply and praised MEPs for taking the lead on AI regulation globally.

“A key goal of our open letter was really to mainstream the conversation about policymaking and regulation,” he said.

“It feels like our open letter really struck a nerve. A lot of people had an urge to discuss these things, but held back for fear of being seen as paranoid Luddites. The most important thing is that this conversation is happening.”

Last week China's cyberspace regulator unveiled draft measures for managing generative AI services, saying it wants firms to submit security assessments to authorities before they launch their offerings to the public.

The Biden administration has also been seeking public comments on potential accountability measures for AI systems as questions loom about their impact on national security and education.

The European Commission proposed the draft rules for an AI Act nearly two years ago under which AI tools are expected to be classified according to their perceived level of risk, from low to unacceptable.

A parliamentary committee is debating the 108-page bill and hoping to reach a common position by April 26, according to two sources familiar with the matter.