For Liberty Vittert Capito, a data science professor at Washington University in St. Louis, the future is terrifying—not because of what AI might do someday, but because of what it is already doing, and the failure of lawmakers to grasp the speed of change.
In a recent conversation with her father—a retired entrepreneur she describes as the smartest person she knows—she was struck by his shock upon learning that AI is not a lab experiment but a technology already embedded in daily life, from Siri to advanced systems. That ignorance, she argues, mirrors the complacency of most legislators, who have the power to act but lack the understanding.
Capito defines AI as an autonomous system that perceives and acts within its environment to achieve a goal. She distinguishes between narrow AI, which has existed for decades and excels at single tasks like chess, and general AI, which reasons and adapts like HAL from '2001: A Space Odyssey.' General AI, once thought decades away, is now imminent, with capability doubling every six to 12 months—a pace she calls a cliff's edge, not a linear progression.
Two immediate threats demand attention. The first is mass unemployment of a scale never seen. Capito predicts that within five years, one lawyer will do the work of five, and entry-level jobs in customer service, data entry, and administration will see a 20-to-1 reduction. Even her own profession faces a 15-to-1 cut. This is not a recession, she warns, but a structural dismantling of the workforce.
The second threat is cybersecurity, with personal consequences. Capito cites Anthropic's delay of an AI model update after it demonstrated dangerous hacking abilities that could crack banking infrastructure. She asks readers to imagine waking up to an emptied bank account or retirement fund, with passwords rendered useless and investigations taking months. Broader risks include attacks on the power grid, military systems, and critical infrastructure.
Capito draws a stark comparison: after Thomas Edison's lightbulb patent in 1880, it took 80 years for electricity to reach every U.S. home. ChatGPT reached 100 million users in two months. Unlike past transformative technologies, AI offers no time to adapt, regulate, or absorb the shock.
She urges legislators to act immediately—not study, not form committees, not schedule hearings for next quarter. They are the only ones who can still intervene. The question is whether they will understand the problem before it is too late.
Capito's warning echoes broader concerns about federal inaction. For instance, the U.S. drug strategy has been criticized as 'flying blind' against new synthetic threats, while experts warn that federal inaction on sports betting fuels a youth addiction crisis. Similarly, Anthropic's restricted AI model triggered an urgent federal response over cybersecurity threats, highlighting the need for proactive governance.
