TV

00:00 "You know, you can't outsource accountability if you rely on AI." 00:04 "It can't give you back the sleepless nights." [Music] Hello and welcome to Insurance Business TV for our second Professional Risks Power Panel of 2026. We're going to look at what the insurance market is doing right now, including the external forces reshaping the risk landscape, and where traditional PI products are facing challenges such as AI. 00:13 Our experts are back. Nick Bezwick is Head of Professional Indemnity in Australia and New Zealand for Newline Australia Insurance, and Austin Rosier is Principal Risk Advisor for OmniShore. 00:39 Welcome back, gentlemen. Thanks for joining us again. 00:46 Nick, let's start with you. If you were to look across professional risks today, what external forces are most reshaping the risk landscape from where you sit? Regulation, enforcement, AI—what does it look like from your perspective? 01:08 Sure. I think the professional risk market has always had to respond to a shifting environment. 01:15 For example, state governments have recently introduced changes to laws affecting the construction sector, and insurers have had to respond accordingly. 01:24 Historically, the insurance sector hasn't really been driven by significant or rapid technological change, but AI is starting to challenge that. 01:33 The use of AI across a wide range of consulting firms is expanding rapidly. 01:42 This introduces new challenges for insurers when considering how firms are implementing the technology. 01:50 These challenges can have a silver lining. Like other sectors, AI offers insurers opportunities to improve efficiency in operations. 01:58 That includes underwriting, higher-quality data management and analysis, and reporting. 02:05 Thanks, Nick. Austin, what external forces are you seeing? 02:14 I tend to agree with Nick that insurance is generally a reactive industry, and AI is moving much faster than the typical reaction times we see in insurance. 02:22 From my side, it'll be really interesting to see how AI gets used in underwriting. 02:30 For example, AI tools might take on a large portion of slip reviews. 02:38 Similar to how AI is used to review thousands of résumés, you could see the same thing happen with insurance slips. 02:47 You might receive declines that are algorithm-generated rather than true underwriting decisions. 02:56 I think there's definitely going to be movement in that area over the coming years. 03:05 It will be exciting, but it will also come with challenges that insurers are already thinking deeply about. 03:14 On the broking side, there are significant and ongoing regulatory changes at the moment. 03:21 That's a sector-wide concern and not limited to professional risks or professional lines. 03:29 We want to ensure we're compliant while still delivering the best possible service to our clients. 03:38 As you both say, we're in a reactive industry. Because of these external forces, are you seeing changes and rethinking around coverage yet? 03:45 Austin, let's start with you. 03:51 I would say it's still quite limited. 03:58 We haven't seen an AI-specific endorsement or standard-form wording emerge yet. 04:06 It's not like cyber endorsements that started appearing on PI policies five or more years ago. 04:14 We haven't seen that sort of development with AI yet. 04:20 I think we'll eventually need to see some policy wording changes. 04:28 Many businesses that were previously viewed as professional liability risks are now using software or AI-driven service models. 04:35 What historically looked like a professional indemnity risk may now need to be viewed as an IT risk. 04:42 That's something everyone will need to be mindful of. 04:51 Brokers need to ensure coverage structures remain appropriate and have the right conversations with insurers. 04:58 We've also seen underwriting algorithms needing adjustment. 05:06 Some products may need to be tweaked or even fundamentally rethought. 05:13 One example is Commercial Legal Protection (CLP) cover. 05:20 We've seen significant increases in claim notifications that appear to have been written by ChatGPT, Copilot, or similar AI models. 05:30 You can often tell from the style, formatting, and use of em dashes. 05:40 These notifications are increasingly being assisted by AI-generated content. 05:47 Sometimes that content is based on incorrect advice from an AI model. 05:55 We recently had a client respond to legal demands and receive a very long reply within a minute and a half. 06:02 It was obvious that the original email had been pasted directly into an AI tool and an automated response generated. 06:10 I think that's going to significantly impact how claims are handled. 06:20 It may also affect how excesses operate and how insurers respond to legal demands if AI makes such demands more accessible. 06:33 Thank you, Austin. Nick, as an underwriter, how are you rethinking coverage, and is AI playing a role? 06:42 At this stage, possibly not. 06:49 When things change, insurers need to rebalance pricing and product offerings across affected areas. 06:56 We're not necessarily impacting the availability of cover. 07:03 As AI becomes more prevalent, insurers will focus on how the technology is managed and implemented by professional firms. 07:12 Particular attention will be paid to privacy, confidentiality, and ensuring AI-generated reports are properly verified before being released to clients. 07:21 There's also a claims impact that needs to be considered. 07:29 It's very doubtful that reliance on AI alone will be a successful defence to a PI claim. 07:36 You can't outsource accountability to AI. 07:44 There needs to be an appropriately qualified and experienced human involved whenever advice is provided to a client. 07:53 So far we've focused on existing professions adapting to change. 08:00 But we're also seeing entirely new professions and business models emerge. 08:09 For example, fractional CFOs and CIOs working across multiple businesses. 08:17 Presumably that's creating challenges for traditional professional indemnity policies. 08:25 Nick, is it difficult to keep pace with those changes? 08:31 Yes and no. 08:42 You need to ensure the right people are in the right roles and are qualified to provide the advice they're giving. 08:55 What's really changing is how technology is used and how that may influence claims outcomes. 09:04 At the end of the day, a PI policy is still responding to the same core exposure: a breach of professional duty. 09:11 It's more about how services are delivered using technology than the technology itself. 09:18 I see AI primarily as a tool that needs to be managed and used wisely. 09:26 The core exposure remains the same. 09:35 Austin, from a broker's perspective, are you seeing situations where traditional indemnity covers are struggling to keep pace? 09:46 Yes, but often it comes down to understanding what a business is actually doing. 09:54 What companies tell investors can be very different from how risks are presented to insurers. 10:04 That doesn't necessarily mean there's a difference in reality. 10:13 Many businesses market themselves as highly innovative technology companies. 10:20 But ultimately, there's still software underpinning the service. 10:27 And usually, there's still a human directing the process. 10:36 A lot of our conversations focus on understanding the real activities and operations of the business. 10:43 I agree with Nick that what businesses are doing often hasn't changed significantly. 10:50 The bigger change is how they're doing it. 10:59 That may require reconsidering which category of risk the business falls into. 11:07 For example, does it remain purely a PI risk, or is there now an IT or cyber component? 11:16 We also spend a lot of time advising clients about AI governance and controls. 11:24 Many AI models don't provide data shielding by default. 11:31 Clients need to understand what data they're sharing. 11:39 A well-known example involved Samsung engineers uploading source code into an AI system. 11:46 That code then became part of the model's training data. 11:55 As a result, Samsung's source code became accessible through the system. 12:03 That's obviously a significant issue. 12:10 Businesses need to understand both the opportunities and the risks. 12:23 Many problems arise simply from a lack of awareness. 12:32 Thanks, Austin. Let's look ahead. 12:39 Where do you see the strongest opportunities for clients to improve their professional risk outcomes over the next 12 months? 12:54 Austin, let's start with you. 13:01 There's a lot of opportunity in the AI space. 13:09 Almost every industry can use AI to improve risk management. 13:17 It can help analyse processes, identify cyber security gaps, and improve operations. 13:25 As much as we all value insurance, it can't solve every consequence of a claim. 13:32 It can't give you back the sleepless nights. 13:39 It can't remove the distraction from onboarding new clients. 13:47 Insurance can cover financial costs, but it can't cover the human cost. 13:55 [Laughter] 14:04 We encourage clients to actively engage with industry-specific technologies and opportunities. 14:12 For example, we're using a new compliance platform called Curium. 14:19 The software helps identify complaints, issues, and potential compliance breaches. 14:27 It can even reference the relevant code requirements. 14:32 That helps ensure we're compliant and doing things correctly. 14:39 Tools like that can help minimise professional indemnity risks. 14:47 Lawyers, engineers, and many other professionals can benefit from similar technologies. 14:54 These tools should be viewed as ways to better protect businesses. 15:02 Of course, they should be used with appropriate caution and oversight. 15:11 Nick, you get the final word. What practical steps would you like insureds to take over the next 12 months? 15:18 The more things change, the more they stay the same. 15:26 Insureds should focus on providing thorough, well-thought-out underwriting submissions. 15:33 That includes clear claims histories and notification registers. 15:40 Risk management procedures should be explained in simple terms. 15:48 Don't try to make activities sound more impressive than they really are. 15:57 And don't misrepresent your business activities, as that could jeopardise a future claim. 16:04 AI will become woven into all of these processes. 16:13 It can improve data quality and make submissions easier to complete. 16:22 AI should be used to improve submissions rather than reduce the amount of information provided. 16:30 That's particularly important in a soft market where some insureds may feel minimal information is sufficient. 16:38 For higher-risk activities and sectors, quality submissions remain critical. 16:47 They're just as important in a soft market as they were in a hard market. 16:57 AI should simply make achieving that easier and more thorough. 17:00 Great. Thanks very much, Nick, and thank you, Austin. And thanks for watching Insurance Business TV. Bye for now.