Diving Deep into Artificial Intelligence with Katja Grace: Part Two

In Part Two of our Artficial Intelligence interview with expert Katja Grace, we talk employment, income and robot babies.

Continuing where we left off last week, in the second part of the interview with Katja Grace we talk about the potential job losses due to AI, her views on minimum guaranteed income, her research areas and her growing up years. In addition, she has some very pertinent advice for young people.

 

AI and Jobs

A study by McKinsey Global Institute has forecast that AI could replace up to 30 percent of the world’s workforce as early as 2030. What are your views on this?

I don’t know a lot about the study in question, but I think such metrics are easy to misunderstand.

Throughout history automation has replaced most people in their jobs many times over, but they are not then out of work—they go on to do the next thing that machines aren’t doing.

Almost everyone used to work in agriculture. Machines do almost all of the farm work now, but people went on to do other things.

All sorts of jobs people had even a 100 years ago don’t exist. So the workforce being replaced is nothing new, and says very little about unemployment.

Eventually, I expect AI to become so much better than humans at everything that it is hard to sell labour as a human at a rate that is worth working for, and so eventually most people will not be employed.

The big question is when that old pattern of work being automated and people moving to more sophisticated tasks will change into a pattern of work being automated and people moving to unemployment.

I have not heard of any sign of this beginning yet, but I haven’t been attending closely to this literature lately.

Off the top of my head, 30% being replaced by AI – even to go on to other work rather than being entirely replaced – sounds high, but these things are very uncertain.

 

AI and Income

If millions of jobs are to be lost to AI what are your views on a minimum guaranteed income?

If things are to go well, it seems important that before selling labour ceases to be a human option, pretty much everyone has other reliable sources of income lined up.

Hopefully under such circumstances a lot of wealth will be being created by superhumanly productive and creative machines.

But it’s hard to imagine everyone having a share in this without large organised efforts at redistribution.

I can imagine various ways of this happening, and have not studied the pros and cons of different policies, but a minimum guaranteed income sounds like a natural and positive way to do this. 

 

Katja’s Research

Please tell us about your research areas.

I am currently investigating whether we are likely to end up with AI that is like ‘agents’ pursuing objectives in the world fairly independently (like the automated corporations I imagined above), or whether most of the ‘thinking’ will be happening in software like Google Maps, which seems less dangerous.

My organisation does a very wide range of research.

Image from iOS 10
AI Expert: Katja Page

For instance in the past, I have learned about the history of cotton processing in the industrial revolution, as input into a project on how smoothly technology generally progresses (and so whether we should expect AI progress to proceed smoothly, or with surprising jumps) and I have run a survey of artificial intelligence experts worldwide, to find out when they think human-level AI is likely to happen, and whether they think it is dangerous (they mostly do).

Tell us about your growing up years and those who motivated you along the way?

I grew up in various wild and rural areas of Tasmania, Australia. I didn’t go to school as much as one might.

My mother usually had an ambitious collection of projects going on, and I think inspired me to look at the world and ask what could be done to make it better, rather than looking at it and asking what slot I could fit into to officially ‘succeed’.

Meeting the LessWrong community and the Effective Altruism community part way through undergrad was a big deal for me, because both are full of people sincerely trying to make the world better, and who are similar to me in their intellectual approach to it.

 

AI Advice

Our readers are mainly the youth in different parts of the world who look up to young achievers such as yourself. A word of advice for them?

I think different people often need different – even opposite – advice, so here are some things that seem good to me, in my circumstances, for which my advice is to just ask whether they might be right for you:

  • Where you put your attention in life can entirely change the game you are playing, even while your physical surroundings are the same. What you habitually do with your attention can make a huge difference.
  • Consider writing a blog. For me this has been a great way to interact with people interested in similar things, to try out ideas fast, and to make something fun and quick at home that other people care about. (My blog.)
  • Try to understand the world yourself rather than deferring to others about what needs doing.
  • If there are things it looks like you ‘should’ do, but you are not feeling it, try to understand at a gut level why they are important. Imagination is helpful here: really picture the good outcome, and the bad outcome.
  • ‘Rapid prototype’: if you’re considering doing something big and hard, look around for cheap, fast ways to test out the plan. For instance, once when I was thinking about whether to have children, I rented one of those robot babies that they give high school students, to find out what that aspect of the experience was like. If you think you might be interested in working in a research area, try jumping into it for the weekend, or talking to someone who works there about the details of what they are doing, instead of starting by trying to get into a degree program to study it. You might find that the actual work isn’t as interesting as you thought it was, or that it’s easy and you can make fine progress without the degree.
  • If you are scared of things, think of ‘doing things while scared’ as an awesome skill that you are getting a desirable opportunity to practice, instead of assuming the objective of getting rid of the fear.
  • Focus on doing good work, not on whether you are being appropriately respected for it. The game of doing good work is way more fun than the game of trying to manage your social success, and often if you succeed at doing good work, people will pay attention to it in the end (at least in my limited experience). 
  • Try to spend time around people doing things you admire and care about – lots of social things are contagious.

Katja Grace runs AI Impacts, a small multidisciplinary research group concerned with long-run risks from advanced artificial intelligence. The mission of the group is to answer decision-relevant questions about the future of artificial intelligence. Their 2016 survey of machine learning researchers was the 16th most discussed academic paper in 2017 according to Altmetric. 


Did you miss part one of our interview with Katja? Check it out now:

Diving Deep Into Artificial Intelligence With Katja Grace

Share this post

Interested in co-operating with us?

We are open to co-operation from writers and businesses alike. You can reach us on our email at cooperations@youth-time.eu and we will get back to you as quick as we can.

Where to next?