00:00Well, Dr. Conor McKeown is lecturer in digital media at the University of Stirling.
00:05Conor, thanks for joining us on the program.
00:07What do you think? Will companies use AI less if they don't feel adequately insured?
00:14Yeah, I think that it depends on the industry,
00:16but there is a real chance here of seeing companies being more reticent,
00:22more cautious about using AI and their standard workflow
00:25if it's not really necessary and not making the kind of impact to the business
00:29that they think would warrant the cost.
00:32But more interestingly, I think what we're seeing are really interesting strategies
00:37being developed by companies to continue using AI
00:41in ways that can be sort of catered for within traditional insurance causes.
00:49What changes do you think will see to kind of traditional insurance policies because of AI?
00:56The biggest thing of all is human-in-the-loop style insurance policies.
01:03And this is really exciting to me because it shows that even though we are going to be using AI
01:08tools,
01:09we need to have experts in place that can double-check the kinds of work that AI is doing,
01:15the kinds of things that AI is producing because of the insurance liability,
01:19because it has to be someone's fault.
01:22It can't be a machine's fault.
01:24So we have to get a human-in-the-loop double-checking work.
01:27And I think that that really is good news for everyone.
01:30It means fewer jobs really being completely replaced by AI
01:35and more of us thinking about how we can consolidate humans and AI in the same workplace
01:41because somebody has to be the person that's ultimately responsible
01:45for the actions and decisions that bots are making.
01:48Now, insurers traditionally use historical data when they're calculating risk.
01:52But how do you do that with AI, particularly at the moment when that data just isn't widely available?
01:58Yeah, that's really interesting.
02:00And I've spoken with some insurance providers, actually in Finland of all places,
02:05and they're telling me that their strategy at the moment is to kind of treat AI and AI use
02:11in a similar vein to how traditional software was used.
02:15So if something goes wrong, if there is a policy dispute,
02:20it comes down to who provided the prompt.
02:24In the same way that you would give insurance policies based on who programmed the software,
02:29it comes down to how was this used.
02:32And we're seeing the same thing from companies themselves,
02:35where companies are less branding themselves to be AI forward
02:39and more using AI tools for certain parts of their workflow.
02:45So we're kind of seeing this on two sides,
02:47traditional contracts being rewritten just to make those certain accommodations
02:52and then businesses themselves kind of hedging their bets
02:56and not being as totally upfront about AI use.
02:59Dr. Conor McHugh, thank you very much for those insights.
03:02And it feels like AI insurance is something we're going to be discussing very much in the future.
03:06Back to you, Sally.
Comments