Artificial Intelligence & Society (feat. Francine Descartes) @BBC

Monday night, at the beautiful Broadcasting House in London, the @BBCBlueRoom organised the first of the Artificial Intelligence and Society events. This is a set of initiatives to foster discussion around how the British public broadcasting service can help shaping the future of AI, and how AI can help shaping the future of BBC. An amazing event around fear, democracy, freedom, and justice with, for, and through AI.

What is the role of BBC in the upcoming AI revolution?

Matthew Postgate, CTO of BBC, addressed this question in his opening presentation. Main message:

We should use world class public services to promote responsible* AI, and use responsible AI to improve BBC services.

*Responsible=free, transparent, accessible, fair.

Technology is in the BBC DNA. Even during very hard times. For example, the BBC during World War 2 trained and employed 800 women as engineers to ensure technology advances.

History of the BBC

BBC for responsible AI: foster debate through broadcasting. BBC will be responsible to bring the discussion on the BBC channels and inform people on the pro and the cons of AI.

Responsible AI for BBC: encoding trust into artificial intelligence services. BBC means trust: people watch BBC because they know it will be impartial, fair, accessible. A Bible of editorial guidelines ensures that these values are encoded in every piece of information broadcasted by the BBC. In the AI era, there must be editorial guidelines for AI services. These will enforce transparency, impartiality, universality: similar to what BBC does with real editors, BBC AI services will need to allow for interrogation, provide answers about their rationale, and not just behave as black boxes.

Royal Society’s Report on Machine Learning: how do people feel about AI

Recently, the RS conducted a survey to understand how people feel about machine learning and AI technologies. Main outcomes of this report:

1. Most people are not aware of the term “machine learning”, but they were familiar with the applications.
2. Depending on the applications, people’s view about AI and its potential changes a lot.
3. Depending on the owner of the AI (e.g. companies vs social good institutions) people’s view about AI changes.
4. Main area of concerns of the public: machines can cause harm, replace people, restrict our options

Main Message: We need to be active on how this technology unfolds in order to make this revolution a benefit for everyone.

AI and the law.

In her amazing presentation, Professor @lilianedwards gave an overview of the fundamental notions of legal governance of robots. This is what I understood.

Keep machine bias under control is crucial to ensure transparent AI, and to build robots that respect human rights and fundamental justice. Ensuring transparency and fairness is key to avoid, for example, racist behaviors in AI services, see the story of the AI trained to automatically predict future criminals which ended up targeting black people.

How?

Next year, the new EU Data Protection Regulation will become an official law (and the UK will have to comply to it until Brexit — and possibly beyond :)). Two fundamental rights:

1. Explanation right: users have the right to know the rationale behind the algorithm — e.g. why did Facebook recommend me this post?
2. Privacy by design: systems will need to be built from scratch for privacy protection, privacy-preserving mechanism should be embedded at the source. In this scenario, AI services are inevitably targeted as high risk for privacy protection.

AI, Utopias and Dystopias

In his mind-blowing presentation, @stephenjcave, Philosopher and Director of the Leverhulme Centre for the Future of Intelligence, blended tales, legends, history, and philosophy in a unique overview of AI utopias and dystopias.

First, Francine Descartes, or how the same technology can mean love for some people, and fear for others. Renée Descartes daughter tragically died at the age of 5. Here, the legend starts. Desperated, Renée built an automaton-version of Francine: a child-robot he could hug, love and talk to, like a real daugther. 10 years later, he embarked with his Francine on a ship to Sweden, to visit the Queen. Scourged by the inclement weather, the sailors started believing that the source of such bad luck lied in those strange voices coming from Renée’s room. One night, they crashed into Descartes room, opened a wardrobe, and found the robot Francine. They were so scared by this devil, this monster that they immediately took her and threw her to the sea. Renée died few months later, in Sweden. Utopias and Distopias: Renee’s love for the robot, versus the sailors’ fear.

People both admire and hate AI because Robots are in between humans and gods. Robots are very human: they look like us, they are build to replicates our mind and capabilities. At the same time, Robots are legendary creatures, very similar to gods and their potential power and capriciousness. AIs are emotional imaginary constructs,

Utopias and Distopias: Power VS Intellectual Dominance.

• Utopia: machines will give us more power (especially military power), and we like power.
• Dystopia: at the same time, we see machines as diabolic ex-machinas: by nature, since Plato’s intellectual meritocracy, we have the tendency to think that intelligence brings dominance: that is why the colonialism and other forms of ruling over the less educated happened.

Utopias and Distopias: Desire VS Alienation.

• Utopia: emphatic machines will allow us to live at full our relationships without all the complexity of being human: machine will not want anything in return for a deep relationship.
• Distopia: robots are human dopplegangers, relationships with machines will bring alienation, and we won’t need humans anymore.