Doug Scott on AI

Archive note: This interview is retained at its original path so links and the historical record are not lost. It is lightly copy-edited for punctuation and readability, while the original HTML is held in the dated site archive. It reflects the author’s views at the time; the current evidence site has a separate methodology and corrections policy.

1. Since you set this site up, what is your interest in AI?

I find it both fascinating and scary that humans are on the edge of creating something that is more powerful than themselves in so many ways (be it 20 years or 500 years, it is close). Does AI become the new species that rules the universe forever, or have other species been here before and something happened? I can’t believe in all of time and with the size of the universe a species has not got to this point before. What happened, or what is happening that we are not aware of or can’t even comprehend?

2. Why should we care about the ethics in AI?

The rules we set for AI, if done well, can create the future universe we imagine. If done well then we could help create something very special; if done badly…

3. Since you have been so active in startups, what is the best use of AI in a startup or business you have seen and why did you like it?

Many people are using AI to do a function better than a human can do. That is another type of hammer to hit a nail, but it does not change the game. The things I like are things that create platforms for game-changing advancements. I am a great believer in non-centralised control and data for the benefit of humanity, and hence utilising AI and blockchain is something I find interesting. People like the teams at Ocean Protocol or Raven Protocol are interesting.

4. What scares you about AI?

Getting the ethics at the beginning wrong. My concern is that we are rushing headlong into advancing the technology without enough thought going into the ethics that the new AI systems will be based on.

5. If a self-driving car had to choose one of two separate people to run over, who should it choose?

You :) On a serious note, we need to realise that all life is not equal. That sounds awful, but humans make decisions every day that have huge implications for other humans. If a human were driving the car and had to choose who to run over, they would make decisions based on certain morals and facts. We need to document these and make sure we make the car able to react based on our morals. We may do this at an individual level or a society one.

6. If you could conceptualise any use of AI that would most benefit humanity, what would it be?

Helping us advance our thinking on integrity, morals and ethics.

7. Who do you admire the most when it comes to AI?

The people I have watched who seem to have some grasp on the enormity are Jaan Tallinn and Nick Bostrom.

8. You have one question you can ask someone about AI. Who would it be and what would you ask?

Donald Trump, Vladimir Putin and Xi Jinping: can you all spend four weeks together, 24 hours per day, in the company of 10 of the most knowledgeable people in AI with no one else around?