Al augmentation or liability amplification? Navigating professional risk

Business Insurance
4 followers
7 Video Views·Jan 11, 2026
On this episode of Insurance Business TV, we explore how artificial intelligence is reshaping professional liability risk. Dennis O’Connell, SVP, Professional Liability at Westfield Specialty, explains where AI is genuinely increasing exposure, where concerns may be overblown, and how responsible use of AI can actually reduce risk.
Learn how governance, data quality, and oversight impact AI-related claims, and what insurers and brokers should focus on as clients integrate AI into their workflows.
Hello, everyone. Welcome to Insurance Business TV as we delve into the professional liability
risk landscape. Like so many aspects of insurance, it is a sector that has been
greatly impacted by the huge rise in the use of artificial intelligence.
But what does that mean for policies? And indeed, what does that mean for brokers?
Well, here to talk us through it all, we welcome Dennis O'Connell,
SVP Professional Liability at Westfield Specialty. Dennis, welcome to Insurance Business TV.
Hi, thanks for having me.
So, Dennis, we've seen recent changes in terms and conditions that suggest that AI
is increasing professional liability risk. So in your view, what are the most misunderstood
aspects of AI risk in the professional liability space? And indeed, how should
insurers approach companies that are actually proactively and responsibly integrating AI
into their workflows? First, I'd say that AI isn't just a
source of exposure. It can actually be a tool for managing and even lowering existing risk.
I think there's a tendency right now to overreact to AI in the context of professional liability
exposure, not in terms of its transformative impact. That's undeniable. But I think we need
to be thoughtful and not jump to the conclusion that this is going to be an Armageddon moment
for liability risk. There are really two key reasons for that. First, most AI-related claims
stem from misuse, weak governance, or poor data quality. And these are fundamentally human errors,
not flaws in the technology itself. And they can be mitigated with strong oversight and clear
procedures. Second, when AI is implemented responsibly, I actually think it can be a
powerful risk mitigant. It can reduce human error, it can improve decision-making,
and it can strengthen compliance. With all that said, I believe insurers are smart to be studying
this issue and introducing safeguards, and to be underwriting this exposure to understand
which insurers are introducing technology responsibly. But the reality is that this
risk landscape is evolving, and it's probably too early to declare what the net effect of AI will be
with respect to overall exposure on insureds that are using these tools.
Yeah, you've given us a lot of the positives of AI there. But if you don't mind, I'll zoom in on
some of those risks. There are, of course, significant risks when it comes to the use of
AI and automation, such as lack of supervision, regularity compliance, and privacy concerns.
So what stands out for you?
First, I agree with all the concerns you mentioned, and I'll add a bit to that.
You started with lack of supervision, and obviously that's a big issue. AI is not perfect.
It hallucinates. And we've seen real-world examples of this, such as law firms that got
themselves in trouble submitting briefs with fabricated citations. And as you mentioned,
those types of mistakes often lead to fines and penalties, or have other serious legal
ramifications. And I'll just add a few other things that I'm observing.
The first is particularly relevant to professional service firms. And the leaders of these firms are
aware and trying to solve for this issue, but it's a lack of workforce development.
We are eroding the apprenticeship model. Traditionally, junior employees learned by
performing foundational tasks. And now with AI doing a lot of that work, future talent may lack
the hands-on experience that's needed to build expertise in their fields.
And finally, I'll add one more thing that I think is particularly scary for insurers.
And that's that AI mistakes can occur more quickly and at scale than human errors,
and they can compound rapidly across systems.
So those are just a few of the significant risks that we're seeing.
Well, okay. Now that I'm worried, we'll flip back to the more positive sides of AI.
So how can AI actually help reduce professional liability losses? Can you share any examples where
AI-driven analytics or automation have demonstrably minimised human error,
reduced risk, or indeed improved fraud detection?
Absolutely. I think we've reached a point where not adopting AI is itself a risk.
As industries evolve, failing to leverage AI can mean falling behind in competitiveness,
or even failing to meet the emerging standard of care expected in many sectors.
And so, as I touched on earlier, there is a benefit that many of our insureds are using AI
tools to reduce their own risk, and by extension, ours. And I'll point out a few areas where I think
we're seeing the biggest measurable difference. The first is fraud detection. Fraud is one of the
most severe sources of claims and professional liability. And AI excels at analysing patterns
and anomalies to detect and prevent fraud. These tools are now widely deployed in financial
services and among our CPA firm clients, where that early detection will significantly reduce
our risk. Another area is error reduction. I mean simply mistakes. Mistakes are the number one
major driver of claims frequency and professional liability.
And AI, when paired with human oversight, can actually review work and catch these errors or
mistakes before they occur. A couple areas where we're seeing this. Many of our insureds will
utilise AI as an additional check on things like contract language to identify harmful,
non-standard terms and conditions, or for compliance. Others will use it to review
designs and do simulations and test their assumptions. But what we're seeing broadly
is really an enhanced decision-making model which reduces our risk. But of course, in all these
cases, we stress there needs to be a human that's involved. Yeah, I think a lot of people like to
use AI, they like to experiment with AI, but they don't necessarily know the best practices. So
what best practices are you seeing among the most sophisticated insureds when it comes to
AI adoption? For example, how important is it to have a formal AI framework, to maintain a human
in the loop as it were, or indeed to provide ongoing employee training? I completely agree
with all of those things you just mentioned. I think it's very important to have a formal
framework. We're seeing most of our sophisticated insureds are doing that. That includes things
like we've talked about, having a human in the loop oversight, especially when performing high
impact processes. This ensures accountability and it provides a safeguard against hallucinations.
Like you said, continuous training is also extremely important. This needs to be more
of a cultural shift and sophisticated insureds are providing training, not just on the capabilities,
but some of the limitations of AI that we've been discussing. And I would also add that it's
important to do scenario planning and have contingency measures so that if you have a failure,
you can ensure continuity across your business. And finally, I would add that most sophisticated
insureds have a process of audit or monitoring to detect issues before they become systemic.
Yeah. And Dennis, just to wrap up for us, do you have any advice for brokers as they
prepare the professional liability clients in what is of course a rapidly evolving AI landscape?
I have a lot of advice for my brokers, but I would advise simply,
we recognize that AI tools are no longer optional. They're optimizing efficiency and
quickly becoming table stakes in most competitive industries. And while we're continuing to study
how these technologies impact the risk environment, our view is that insureds
who have implemented strong governance and oversight are well-positioned going forward.
Yeah. Great tips there. And Dennis, huge thanks for your time today. No doubt AI will be the
centerpiece of many more discussions. So keep it right here on Insurance Business TV.
