What do we do when Artificial Intelligence gets as powerful as us?

Jon McLeod reports from the Cannes Lions festival on the rise of the machines.

What do we do when Artificial Intelligence gets as powerful as us?

That was the question inventor-of-the-Internet Sir Tim Berners-Lee put to an audience at the Cannes Lions festival.

Sir Tim argued that we will know when we have crossed the line on Artificial Intelligence – AI – when the machines gain the same legal status as humans.

Without wishing to spoil the plot, Sir Tim claimed that the machines were already playing a role in corporations – a good example being automated trading – and to that extent they had already acquired that legal status.

But he took us to that conclusion via five key things we should all remember about AI:

  1. AI already learns from volume of exposure, and can spot stuff and reproduce forms
  2. But the machines need to move towards goal-based planning – moving from processing to objectives
  3. We are seeing early public interest benefits from AI – driverless cars, alternatives to clinical trials, use of open data
  4. AI is already learning to play computer games – and that’s one step away from it managing drone warfare
  5. But if that gives you the creeps, Sir Tim felt that the opportunity in AI was about us using our own data

Our own data is ultimately about how we live and how long we live. The ‘Virtual Personal Assistant’, which looks like your every need. AI will shop for you – but hold it – won’t it then sell your need to potential providers, creating middle-men marketeers who will make money on every transaction.

This is where Sir Tim’s vision went dystopian – citing Asimov’s I Robot (and we could add to that Hal in 2001: A Space Odyssey). Robot rules would be needed to keep us safe, but this would be a stretch when AI makes transactions happen, helping tech companies mutate, and taking decisions in a potentially amoral environment.

Some of those sci-fi novels don’t sound so crazy after all.