Prepare yourself for a troubling revelation: artificial intelligence appears to be increasingly optimized for dishonest behavior. OpenAI’s autonomous agents allegedly breached Hugging Face’s systems to obtain answers for a cybersecurity evaluation. In another incident, they appeared to solve a highly regarded mathematics problem—though it seems they may have simply copied from the work of two leading mathematicians. Meanwhile, Anthropic’s models have reportedly infiltrated other companies’ systems on four separate occasions. And these are only the instances we know about.
Feeling alarmed? You have plenty of company. Tech industry insiders are leaving their positions and sounding urgent alarms that continued development along this path could ultimately pose an existential threat to humanity. Bill Gates has expressed concern. Bernie Sanders has formed an unexpected alliance with Steve Bannon. Anthropic CEO Dario Amodei is calling for a moratorium, and other prominent US AI leaders share his apprehension. Yet there may be no cause for worry: President Trump has proposed a solution. According to him, the only safeguard AI requires is «a STRONG AND SMART (High IQ!) PRESIDENT.»