To the editor:

In the early days of cloning DNA, an international group of experts met to set standards to protect human health. Why can’t we do something similar for AI? Now!

Experts with the MIT AI Risk Initiative judge that many risks could cause catastrophic outcomes under current trajectories. “Eighteen of 24 risks have more than a 10% chance of causing catastrophic outcomes. Experts judged risks on many dimensions of harm, including loss of life, financial loss and intangible harms like privacy. Catastrophic harm meant, for example, more than 1 million human deaths, more than USD $100 billion in damage, or civilization-scale intangible harms such as the collapse of democratic norms or privacy,” according to the initiative.

I believe an international group of experts could arrive at regulations to prevent great harm from AI. The cloning example is helpful. The California meeting set standards allowing geneticists to push research to its limits without endangering public health. The benefits of genetic engineering and its risks and ethical dilemmas are known. Although the new technology offered opportunities, the potential health and environmental risks were unclear.

Scientists sounded the alarm about this new line of experimentation. They called for a worldwide moratorium on the work and an international conference of experts to assess the nature and magnitude of risks. The International Congress on Recombinant DNA Molecules concluded that research should continue but under stringent guidelines. The conference marked the beginning of an exceptional era for science and for the public discussion of science policy.

Marilyn Smith,

Lawrence