Time and time again we have created artificial intelligence (AI) systems to help solve our problems, but what happens when the AI systems become the problem?

Artificial Intelligence systems have been created to help humans work faster, respond more justly, manage more and make fewer mistakes, but now the solution has become the issue. As these systems progress and become more prevalent, ethical and existential risks have emerged. Brian Christian argues that it turns out there is only so much AI can do before it becomes painfully clear that humans need humans. We need empathy and connection when determining bail amounts. We need doctors who know our names in order to feel cared for, not just machines that have downloaded our health data. Not everything can be outsourced, but so much already is and it now becomes a dilemma how to rein it in. What happens when our machines outsmart us, or an enemy outsmarts our systems? How do we realign?

Christian investigates these questions and more in his new book, The Alignment Problem: Machine Learning and Human Values. Join us for our conversation about what must change culturally and in the world of tech to ensure that humanity remains our north star.

NOTES

Christian photo by Michael Langan