Opinion: Regulate AI Abuse Without Choking Development

Summary

This opinion essay argues that the United States should continue developing artificial intelligence while regulating specific abuses and dangerous applications. It presents five reasons for caution: sufficiently powerful AI could become difficult to control, enable cybercrime, fraud, deepfakes and biological or military threats, disrupt white-collar employment, concentrate power among companies or governments, and outpace lawmakers and regulators. It then sets out five reasons to accelerate: AI could support work on cancer, Alzheimer’s disease, new drugs, materials, batteries, nuclear power, engineering and mathematics; give professionals digital assistants and raise productivity; make high-level expertise more accessible to students, small companies and researchers; add intellectual capability across medicine, engineering, manufacturing and defense; and strengthen the United States in competition with China. The author describes China as treating AI as economic, scientific and military infrastructure, with uses ranging from cyberwarfare and intelligence analysis to autonomous systems, weapons development, logistics and manufacturing. The essay therefore frames moving too slowly as a national-security, economic, semiconductor, scientific, military and cybersecurity risk alongside the risks of moving too quickly. Its proposed policy distinction is to prosecute AI-enabled fraud and cyberattacks, regulate weapons that use AI, and restrict dangerous biological assistance, while avoiding a broad conclusion that producing more intelligence is itself the problem. The author uses traffic laws, aviation safety and cybercrime enforcement as analogies for controlling harmful uses without banning the underlying technologies. The conclusion calls for building power plants, chips and data centers, training scientists, securing systems and punishing abuses, while pursuing AI development with safety measures and oversight.