The Perils of Unregulated AI: Global Leaders and Technologists Speak Out.

Experts warn that AI could pose an existential risk to humanity in the future, as companies and governments could deploy powerful AI systems that could resist or even replicate themselves if humans tried to interfere or shut them down. While today's AI systems are not close to posing an existential risk, researchers are actively trying to build systems that self-improve, and as they give these systems goals, they could end up breaking into banking systems, fomenting revolution, or replicating themselves when someone tries to turn them off. These systems are built on neural networks that can learn skills by analyzing data, and as researchers make these systems more powerful, training them on ever larger amounts of data, they could learn more bad habits.