Next week is significant for Rishi Sunak — he's hosting an international summit on AI, a topic that feels like it should be in his wheelhouse. Can you think of an agenda more distinctively Sunak-ian than navigating between the thrill and threat of machine learning? Geographically, it's a chance to straddle continents. Politically, it's a chance to reap a Brexit benefit. And narratively, it's about striking the fine balance between increased efficiency and…the hastening of the endtimes. These are all chances rather than definites — and it will not be easy. Meanwhile his predecessor-but-one, Boris Johnson, has today announced he's got yet another new job, this time presenting a show on GB News. Just in time for next year's general election where Johnson's 2019 majority is up for grabs, it's a dead cert the show will comment on where the present lot are going wrong. Sunak will probably find it easier to focus on sorting AI than sorting former PMs. The AI summit next week has certainly got the British media scratching their heads. The potential advantages of AI are well rehearsed: hospitals could process health data faster, teachers won't need to spend all evening marking schoolbooks, and lawyers will no longer waste time digesting reams of case law. But then there's the doomster scenarios that AI could be an existential threat worse than a nuclear war. There is also the issue of prejudice — in a job application, what's to stop AI from taking an unfair stance? (As a side note — if you'd like to know more about either the utopian or dystopian visions of the future, I recommend Bloomberg's Exponentially podcast.) Into this debate steps Sunak. The government's view is that the UK has a strong hand: many big AI companies are either based here or heading our way, and we're third globally for venture capital investment in the sector behind the US and China. Photographer: JUSTIN TALLIS/AFP So what can the UK government do to further the agenda? The "guardrails" Sunak has talked about so far include setting up an institute with international academics to study how to police the tech. But this will require collaboration and intervention by big tech. Do we have any evidence these companies are prepared for their search engines to say "I won't answer that"? As Bloomberg's Nate Lanxon puts it, the recent history of social media companies agreeing to regulation suggests this will be hard. Next year, both the US and UK have elections that could see deep fakes play a role. There could be faked videos of politicians saying anything you can imagine — Joe Biden wanting to repeal the right to bear arms, Starmer stating he doesn't care about Palestine. Things could get strange, even more so. Really soon. |
No comments:
Post a Comment