I had an email from a UK high school student, Molly Patterson, who was researching “Are Science and Technology Going too Far?” for a project. She asked some questions, which I thought were interesting. Molly agreed that I could share her questions and my answers on the blog. Here they are.

  • Do I think that the capabilities of technology will ever cease to grow?

Maybe, but it’s hard to imagine what might stop them growing. If we think of technology as the space of all possible ideas, and our capabilities as increasing as we explore that space, then there seems to be so many ideas that it feels like our capability should always grow. But things might go wrong, any threat to our civilisation is also a threat to our technological progress. There are may be physical limitations to many of our capabilities. For example, there are limitations triggered by the fact that the speed of light upper bounds how fast information can travel. Maybe we’ll get to a point, in some domains, where we find new or equivalent limitations. That would also restrain us.

  • What is the purpose in the research and development of AI and what are researchers currently working towards?

Just like any creative process, I don’t think there’s a single purpose for research and development. Of course, people justify investment in research through suggesting it makes society better or that it develops economies. But motivations will vary across individuals. Many people enjoy the intellectual difficulty of the problem. It’s a puzzle, like an incredibly difficult crossword. Some people are motivated by particular aspects of the problem: for example understanding language, or understanding how we perceive the world around us. I think for many people understanding artificial intelligence is really about understanding ourselves. For me, I’m interested in problem solving, and tools from artificial intelligence are often the right way of solving problems. If I’d been alive 160 years ago then I would have liked to have been working on the solutions to their problems, like efficient transport through steam engines or communication via the telegraph.

  • Which industry do you think might benefit the most from the use of AI and why?

I’m most excited by applications in health. That’s because our health is incredibly complex. It is an interaction between sub-microscopic behaviour (where try to carry out interventions by prescribing drugs), and our whole sense of well being. We are getting a lot of data about the state of our health, too much for a human doctor to deal with. But doctors also have a lot of expertise. So it’s very important that any solutions we develop operate in concert with medical professionals. There are lots of challenges though, particularly around privacy and the sensitivity of patient data.

  • Do you believe that AI could pose a threat or a danger to humans, and do robots have the potential to ever replace humans?

It already is posing a danger, just like any new technology. The first passenger railway killed the MP for Liverpool on its first day of opening. A particular challenge is how rapidly we can deploy AI solutions. This means that if there is a design problem then it could effect many people before we realise the extent of the problem. I don’t think it is (yet) posing a threat to humanity though. And I believe that for many of the threats we face, for example climate change and unsustainable energy consumption, artificial intelligence is a key component of the solution.

AI can definitely replace humans in particular tasks. A lot of people are working on autonomous vehicles which would replace people who earn their living as drivers. Other jobs like stock brokers don’t exist any more because they have been replaced with algorithmic trading. In the past many jobs were replaced by more efficient technologies also. Weavers and spinners were replaced by factories in the 18th and 19th centuries. But there are many tasks that are well beyond AI’s current capabilities. For example, yesterday I received my furniture from Ikea for a new apartment. I spent 8 hours assembling the furniture. That gave me a lot of time for thinking about how difficult that would be for a robot to do. Not only the manual dexterity skills to manipulate the spanner and the parts from the box. But following of the instructions, and the use of common sense and experience to fill in gaps in the instructions. A lot of the time while assembling I didn’t have to think intensively, but there were moments when I had to recognise I was going down the wrong path. This is well beyond the capabilities of our current robots and artificial intelligence.

I believe we could obtain robots that could do all of the above, but it is beyond my capabilities to know when, I’d say it’s over 20 years away (which means “I don’t really know”). However, robots can never replace humans just as a flower can’t be replaced, nor can the patterns on a butterfly’s wings. There is a beauty at the core of humanity, the way we communicate, our art, our music. A beauty that is quintessentially human. Computers can’t replace our particular intelligence, and the way we deploy it, because our intelligence is also the consequence of our limitations, many of which computers don’t have. In particular we are limited by the bandwidth with which we can communicate with one another. Computers communicate with much higher bandwidth, removing one of our limitations. As a result I believe a computer can only emulate our intelligence not truly replace it.

  • Overall, do you think that AI development has more positive or negative effects on society?

I believe it has more positive than negative effects, but we shouldn’t be naive about what can go wrong, otherwise the balance will quickly shift from the positive to the negative.