Good AI, and Dangerous AI ... walk into a bar
what can go wrong, mixing alcohol, strong opinions and AI
Geek joke alert!
Good AI and Dangerous AI walk into a bar; they order a binary cocktail and start a debate, but the bartender quickly realizes it's time to reboot before they crash the system.
Ok Let’s keep this civilized, no need to get nasty about my sense of humour. The debate has a lot of people talking. Do we need to regulate AI? Is it going to kill us all? or produce so much dis-information that we won’t know true from false anymore.
What about the good side of AI and the good it will have for society? Education and healthcare could use a boost to help us all, my two cents.
Context on AI and where it’s going
Check out this presentation from Greg Brockman from OpenAI showing off the future of ChatGPT integrated with other services. Spoiler, not a spoiler - little imagination required, AI will be doing lots of stuff for us!
The good AI
Check out this presentation by Khan academy founder on how AI can be a super tutor to kids, and a super teacher’s assistant. Getting everyone educated to their full potential, has never been more real with AI. It has two benefits. First everyone is contributing to their full potential regardless of where they were born. Second, education is how we will stop mis, dis - information.
The chart below is a powerful reminder that more 1-1 help can create tremendous results in education. The problem we have had so far is we could not scale 1-1 to everyone, with AI we have a real shot. This is a game changer.
The bad AI
In the short term, Dis-information with AI is likely. About dis-information I think it’s a lot more likely to be the opposite. Because people imagine they know what AI can do now, shady folks when caught on tape doing bad things will now saw this is fake news, it was done with AI. So AI will have broad shoulders in addition to generating lots of dis-information.
In the mid-term job losses are also likely. There is no doubt in my mind starting today we will do more with less because of AI. The production of anything white collar jobs are associated with is going to be impacted. Productivity with an AI assistant will raise the bar for everyone and every organisation. Lots of companies will be the dinosaurs facing the ice age regarding adapting to AI. They will go extinct, and this will make a lot of people lose their jobs. Because AI is powerful, maybe this is the optimist in me, it will also create many new types of jobs, jobs we could not think of before AI became what it is today. All in all I can’t imagine being the catastrophe people make it to be.
In the long term we could lose control of it. I predict we will have a doomsday clock and dashboard to keep us alert to the threat. We have been walking the fine line with nuclear, since the 1950ies. I’m afraid, but I’m still hopeful we are smarter than we look when we risk dying ourselves with the weapons we have. Think of chemical weapons, the wind changes from being to your back, to being headwinds, and you are no longer the attacker, but the victim. I believe something similar will hold true for AI.
So What'?
I’m amazed at how AI has captured my imagination, I’m sure it’s the same for you. Bad actors will continue to do terrible things. It’s important we continue so our own AI is not inferior. It’s important we find places for it so it can lift us all up. Education is one, but accessibility is also especially important in healthcare. Getting smart and portable devices in the hands of non-specialist caregivers will be a big plus to the entire world. No questions we need to be careful, we need to ask questions. We need to adapt the regulations when the concrete risk of something is imminent. We can’t do this if we do not have the capability to observe it in action. As Brockman says in his preso - It’s good it is in the open now, so many more people are aware, so many more questions are being asked. It beats the scenario where it’s all done behind closed doors in secrecy.
I’m in the camp that benefits of AI to society will outweigh the risks. What do you think?