news

BACK

Special Report: Does Artificial Intelligence benefit all parts of our society?

20.08.2015

WOMEN’S FORUM SPECIAL REPORTS

An ongoing series of articles highlighting upcoming topics and topical outcomes from Women’s Forum for the Economy and Society events around the world

 

Does Artificial Intelligence benefit all parts of our society?

 

Interview with Joanna J.Bryson, Reader and Head of the Intelligent Systems Research Group, University of Bath, speaker at Women's Forum Global Meeting 2015

 

 

 

Women’s Forum: Artificial Intelligence is a very promising field, what are your biggest challenges today?

 

Technologically, challenges are falling by the day, not because AI is easy, but because so many great minds are being thrown at the problem because we can all see the economic promise, and for that matter the military consequences.  We are less astute at realizing the social consequences.  I think one of the big challenges is helping people globally to understand how much power they have in their lives and communities, and power they have given away, because what we do next is pretty easy to predict from what we or other people like us have done in the past.  This power of prediction allows us not only to sell things but also to organize, for better or for worse.  You can see who is likely to innovate, and whether you think of that as promise or trouble depends on what you stand to gain or lose.

My research right now is on trying to help people understand the benefits – and honestly, the costs – of contributing to public goods.  Normally public goods are thought of as things like schools, parks, co-working spaces, but they also include everything from language and culture to governments or shared housing.  Privacy–and it's opposite- are two example of things we don't normally recognize as public goods, and that may be forever changing without us thinking much about it.

Another challenge I and my students are particularly interested in is Systems AI – how do you take all this clever machine learning, planning etc. and make a system that behaves as you want, and that's users can understand well enough to know how to use it and when to trust it?  This question is called "transparency".  We have been working on the question for decades with the goal of benefiting people building real-time systems like robots and computer game characters, but now we are trying to shift gears and benefit the owners of AI as well as its builders.  A sensible conception of AI makes it not only more useful but also less dangerous.

 

 

Women’s Forum: What are the most important fallouts of AI studies? Beside science and technology, are there other or unexpected fields that can benefit from AI research?

 

Artificial Intelligence affects absolutely every field, absolutely every human endeavor. This is because at the fundamental level, intelligence is intelligence.  Whether it is natural or artificial has very few consequences in terms of computational power.

Artificial Intelligence is an extension of our own intelligence, of our own agency, and this is why it affects every discipline, because it makes every one of us more intelligent.  It make some subset of every domain of research easier, which doesn't mean we should neglect the areas that are not benefited, but does mean that the landscape of academic endeavors is fundamentally changing.  Well, it is always changing due to our learning, it is only that this is happening even and ever faster.

 

Women’s Forum: What does it take for research to progress? Are there any obstacles?

 

I don't think that there are any special obstacles to AI research, in fact there is so much money in it that we are making super fast progress.  Well, but I guess that breeds its own problems, and then there is in fact one special problem.  The special problem affects also areas like psychology and moral philosophy and biological evolution.  When you talk about intelligence, you come too close to people's sense of self, too close to their identity.

People say we have not achieved it or we will never achieve it, and this is not only annoying and sometimes obstructive, now it is dangerous because they don't realize how much the world is changing.  So now back to the other problem of rampant success.  I do worry about how some people who have degrees in physics or careers in engineering or business suddenly see this burgeoning success and make statements that are not sufficiently informed about what AI can and cannot do.  Others dismiss any moral concerns whatsoever, and I think this is also wrong.

But scaremongering or worrying about existential threats to the species I think is going too far and not based on good evidence.  The real existential threat of AI is the philosophical one, it makes us wonder what we are for, if a machine can do the things we can and animals can't?  Are we just animals that are super good at computing?  Yes, in fact we are, but philosophy has been dealing with this fact for centuries.  I think we need to return some attention to the humanities right now.

 

 Women’s Forum: What are your objectives in participating with the Women’s Forum Global Meeting 2015?

 

This will be the first time I'll attend the Women's Forum Global Meeting.  I'm very flattered by the invitation and hope that my talk will be useful to participants. Other than that, I hope to learn, but then I always hope to learn. In particular, I am not often at meetings dominated by women, so I hope I will learn more about gender differences in organisational structure and leadership. Since my research has taken me into public policy, which is not an area I have any special training or education in, I'm also particularly interested in how to be efficacious in that arena.

 

Joanna Bryson

Joanna Bryson has broad academic interests in the structure and utility of intelligence, both natural and artificial. Her current sabbatical at Princeton's Center for Technology Policy is on “Public Goods and Artificial Intelligence”. She holds degrees in Psychology from Chicago and Edinburgh, and in Artificial Intelligence from Edinburgh and MIT.