I’ll just start by talking about a problem that relates to this misunderstanding. The problem is loneliness. We have a massive loneliness epidemic in this country—as well as throughout this planet. South Korea is going to spend a couple billion dollars on loneliness maintenance per year, starting next year. China’s going to follow. America is in third place. We are the loneliest non-Asian country on the planet. The impact? I can give you some basic statistics. 73% of millennials report feeling lonely and desperate. 70% of teens feel excluded when using social media. if you were 40 years old or younger and male during COVID, you are more likely to die from suicide than you were from respiratory complications. Critically, it is important to report that the second leading cause of death for people under 25 in America is suicide. That stat about suicide was a real big one for me. I ended up going to the United Nations and got in trouble there. I was giving a talk on AI for global good at the UN global good conference. I said, who gives a damn if the trees and the whales are fine if all the kids are killing themselves on purpose! We should prioritize the epidemic of suicide.

What does it mean if members of the most educated population in history are killing themselves at a rate that we’ve never seen? It comes down to something very fundamental. It’s a very simple mistake we’ve made with computer systems. We thought social media was going to make us feel social. We thought the internet was going to connect us. And we thought the exchange of ideas was going to democratize everything. All the things that we have seen during the last 12 hours [election of
Donal Trump as President] suggest that this is not the case. What we have are echo chambers. We have silos and we have people feeling more and more disconnected from one another than ever before. And social media is making that worse. There are many studies that we all have referenced. We know that these studies support what I have said. I don’t think that it is controversial to say that the last one and a half technological revolutions have made everything worse.

Now we’re in the midst of a technological revolution that they are promising will be more impactful and revolutionary than the creation of the internet. I agree with them.  AI is really going to be able to do remarkable things. If we say that the last one and a half revolutions really screwed us up, then we must ask: what lessons are we learning from that which apply to AI? What can we do about it? This is my fundamental question and I’m looking forward to discussing it.

All those other systems, call it Facebook, Instagram, TikTok, The List, YouTube, whoever. They all created algorithms, learning systems that just kept paying attention and optimizing for time on site—because that’s obviously directly connected to money for them. Time is money. They wanted to get people to stay on their sites. And that’s just what they told the algorithms to do. Make site visitors stay, let’s figure out how to make them stay. In the process, we accidentally trained a whole bunch of algorithms to optimize dopamine production. And that dopamine production created a whole bunch of slot machines that are shaped like phones and computers. There’s a great book called Addiction by Design. This book is about video poker and video slot machines that were built around the idea of dopamine. The machines make us feel the lack of dopamine and drive usage. That is what we have in all these tools.

TikTok is probably the most impressive algorithm of any of them. TikTok is incredible. So is the fact that a Chinese company has an algorithm that’s designed to be emotionally and psychologically manipulative. The algorithm is pointed at the brain stems of our entire millennial and under generation. This is really scary. And the data supports that addiction is occurring.