Lawmakers should regulate harms, not technologies. That is the argument I make in a point-counterpoint syndicated by the Las Vegas Sun and papers across the country. Excessive caution about a new tool carries costs that never show up in a hearing record, because the people who would have benefited never know what they lost.
Three examples from this year show the pattern. Federal regulators ordered faster grid connections for AI data centers because the United States is competing with China, yet local opposition built on theoretical concerns about noise, water, and power draw keeps delaying infrastructure that serves everyone. Several states banned rent-setting software that analyzes market data, as though the algorithm rather than supply and demand set the price; in Austin, rents fell while the software was in use because the city permitted enough housing. And Nevada and Illinois prohibited AI therapy tools, removing a resource from the person with anxiety at 2 a.m. without addressing any demonstrated harm.
Two questions should come before any restriction. What specific harm are we addressing? Can existing law handle it? Antitrust and consumer protection statutes already reach collusion and deception. Enforce them, match the remedy to a defined harm, and leave the tools alone.
Read the full piece: How strictly should AI be limited? Making the case for a light hand (Las Vegas Sun, August 10, 2026)