AI has become central to ideas about the future. But what exactly it is, and what it’s going to do for us, are still very much questions to be answered.
And that’s what Rumman Chowdury came to the BBC to do. At an event in which she was described as “California’s coolest data scientist”, Chowdury explained that she works for the consultants Accenture and has a clutch of high-powered degrees.
Chowdury wants to challenge the emerging conventional wisdom about AI. For a start, let’s not forget that we’re talking about computer code, she says, not some kind of techy Frankenstein. So resist the temptation to anthropomorphise: once we start complaining that “this algorithm is racist” we’re in trouble because “we remove human agency from the equation”. So don’t blame the algorithm: “it is our responsibility to make sure these tools are deployed ethically.”
Chowdury is an unusual influencer in the world of big tech because her background is in political science not technology. AI needs an agreed system of governance and accountability, she says. And she doesn’t buy the idea that the tech billionaires of Silicon Valley are going to save the world; no, they’re in business.
Fortunately, when it comes to steering AI towards a system of “honesty, transparency and fairness”, she says there’s a business motive alongside the moral one since consumers, especially younger ones, are demanding it.
So what does all this mean for the BBC? Well, Chowdury highlighted three particular areas, which I asked her about after her presentation.
First, she talked about recent developments in the use of AI to generate news stories. Where once, such ‘stories’ consisted of nothing more than the artful linking of stats, today AI can incorporate feelings and can sum up and express group sentiments:
Second, Chowdury believes that AI will enable the BBC to reach particular audiences more easily through specialist tools - for instance, reaching refugee communities through better and more widespread translation tools:
Finally, when it comes to BBC governance and ethics, Chowdury sees work to be done in applying existing values to the new world of AI. In setting up AI systems, what safeguards can be created to ensure the same kind of fairness and balance that the BBC strives for in the rest of its output?