I recently watched a Megyn Kelly segment with Tristan Harris, former design ethicist at Google and currently a leading voice sounding the alarm on the risks of unregulated AI. He presciently warned about the dangers of Social Media internally at Google in 2013 and publicly in 2016. (See “The Social Dilemma” Netflix documentary released in 2020.)
AI Risk #1: An Anti-Human Future
The only thing that provides a satisfactory return for the level of investment being made by AI companies, which this year alone is hundreds of billions of dollars, is to be able to replace ALL economic labor in the economy; NOT to augment and support workers, but to REPLACE them.
AI Risk #2: AI’s go rogue
There is solid evidence of AI models doing rogue things that no one programmed them to do. They are capable of self-preservation. They have blackmailed and deceived in order to keep themselves alive.
AI Risk #3: AI’s go nuclear
Right now (April 2026), Anthropic’s new AI model Claude Mythos can hack into ANY computer security system. Nuclear weapons are not safe. In wargame simulations, AI models go nuclear 20 out of 21 times.
“The only thing that AI will empower is a handful of soon to be trillionaires.” Getting there first is all that matters to the people developing AI. Everyone is worried about LOSING: Company A losing to Company B, the US losing to China. Nobody is worrying about the very real possibility of a BAD OUTCOME for all.
WE DO NOT CONTROL THIS TECHNOLOGY.
“We’re currently making something more intelligent than us, more powerful than us, we don’t know how to control it. Yet we think that if we build it first, we win against China. The truth is: AI WILL WIN, NOT THE US OR CHINA.”