Panelists called for a range of protective and preventative remedies for what they described as the coming tsunami of AI, including requiring more data transparency from tech companies by enabling regulatory agencies and researchers to get more access to the information they’re holding, using consumer-protection laws to ensure that AI models have been rigorously tested for potential negative outcomes, adopting the European model of not allowing AI companies to deploy their platforms until they have proved they are safe, or even instituting criminal penalties for AI companies who unleash unsafe technologies. “As a former prosecutor, I can tell you one thing,” said New Mexico Attorney General Raúl Torrez, who earlier this month won a $25 billion judgment against Meta. “When you bring out handcuffs, suddenly everything changes.”

Not all the issues around AI can be solved by the law, however. One concern is the socialization of very young children by AI, including via voice-activated assistants like Alexa or very agreeable AI-enabled toys. “What do kids start to learn to expect from social practice? Do they expect someone to always agree with them? Is it always easy? Is it always like a transaction?” asked Dana Suskind, an early-childhood expert and professor at the University of Chicago. Human relationships are a lot messier, she noted. “They're chaotic, and that sort of mess is where a lot of learning happens.” Parents, who were caught off-guard by the influence of social media and digital devices, need to be educated about AI and encouraged that their imperfect parenting, as well as other human interaction, is vastly more beneficial to children than an AI companion, she said.