Good morning,
Artificial intelligence is going rogue. Last month, OpenAI admitted that it lost control of an AI model, which went on to hack another company—the first known fully autonomous AI cyberattack. Anthropic has since made a similar disclosure, adding to the pressure on governments to introduce regulation to prevent runaway AI or other disasters.
Yet U.S. policymakers are in a bind, argued Sebastian Elbaum and Sebastian Mallaby earlier this year. It is nearly impossible to simultaneously prevent a national security catastrophe, spur the innovation necessary to compete with China and to generate economic growth, and allay the fears of an anxious public. After weighing the tradeoffs of different regulatory approaches, Elbaum and Mallaby identify the best way out of what they call the “AI trilemma.” The technology is far too powerful for governments to do nothing, they warn. AI “portends social and psychological upheaval on a scale at least equivalent to the Industrial Revolution”—an event that “set the stage for a century of political revolutions and world wars.”
Until next week, |
