House of Lords to probe AI data slurping
Who owns all that valuable training info?
How technology giants own and use your data will be a focus for our noble and learned friends on the new House of Lords Select Committee on Artificial Intelligence.
The Parliamentary committee yesterday announced the UK's first public inquiry "to consider the economic, ethical and social implications of advances in artificial intelligence".
Interestingly, among the areas on which the committee is inviting evidence, we spotted...
The data-based monopolies of some large corporations.
The concern isn't new. One of the few times Parliament has discussed AI was in a debate on "New Technologies" in 1985, when Nicholas Fairfax, the Lord Fairfax of Cameron, a peerage in Scotland, worried about:
... the serious consequences for the individual in the concentration of information, and the consequent problems of protecting the confidentiality of that information. One of the most significant developments of this new technological revolution is of course the ability to centralise, collate, and obtain access to massive amounts of information.
He also raised possible conflicts of interest for firms using AI. But it's the issue of user-generated training data that the Committee will want to examine. The bulk collection of this data is now routine – think of your Android location data – and apparently vital for the learning algorithms to work. But large tech companies don't want to acknowledge you as the owner.
Tim Clement Jones, former Liberal Democrat president, called for ethics in AI work earlier this year. He's also aware of the need for personal property rights, helping to single handedly reform the 2003 Licensing Act, which had criminalised the use of a trumpet in public.
The committee includes Labour lords Clive Hollick and Dame Joan Bakewell, and "Third Way" guru Anthony Giddens, director of the LSE, the Lord Bishop of Oxford, and science writer and columnist Matt Ridley.
You can see the others and get involved here. ®