And their off! The old racing call seems like a good fit for the first week of a new business year, as everything and everyone that was put to the side for the last two weeks of December comes racing back onto the desktop.

And so it has been as well with articles, blog posts, videos, podcasts and the like this week, as everyone fires up the old content generator again.

For anyone new to this series of posts, here’s a rundown. My “Five Favorites for Friday” blog post is a weekly series – published, unsurprisingly given the title, every Friday – reviewing five articles, videos, podcasts or the like that caught my eye during the preceding week and that I want to discuss and share.

And with that, here is this week’s fabulous fivesome:

1. Insurance is morphing away from being simply payment for a covered loss and to a form of partnership, in many instances, between insurer and insured to both first avoid and mitigate loss, and then to reimburse for it if needed. This will be no surprise to many in the industry, but I think for others, they may not have noticed how much of this has occurred, simply because it has taken place over the course of a number of years. If you haven’t been paying attention along the way, it can seem, to steal an old phrase, like the changes occurred slowly, then all at once, but in reality, it has been a long and gradual process, but one that now seems to have hit the proverbial tipping point. I could tell a host of stories about this, such as what EPLI coverage has changed into since I worked as a coverage lawyer on its first rollouts, or the extent to which the scope of both coverage and insurer control has changed in fiduciary liability policies from the time I first defended a case under one until today. This short article from Sompo on the risk mitigation and prevention aspects of cyber policies illustrates the point nicely. With climate change and tech risks becoming ever more central to the insurance environment, this type of faux or quasi partnership between insured and insurer either is already or will be the wave of the future.
2. AI in claims handling is a fascinating issue on a number of levels. First off, there is no question that some aspects of claims handling are ripe for AI intervention regardless of the type of coverage at issue, and that some forms of coverage – first party in particular – are well suited to its widespread adoption. On that last point, for instance, there is no reason, beyond the challenges of the software, that AI can’t determine whether a cracked car window claim is covered, apply the deductible and approve – or reject – a particular vendor. The issue is going to be more in third party coverages and similar areas where reasonable people can disagree on coverage or settlement decisions made by an insurer. I have written about this in the context of claims handling statutes and bad faith claims more than once, such as here. This Fenwick & West piece, and this Lion Specialty post, among many others, make clear that insurance regulators plan to tackle this directly and publicly. Barring federal preemption, I don’t think there is any doubt that state regulators and legislators are going to assert control over this issue and dictate some of the key terms of use of AI in insurance, at least in claims and in policyholder facing aspects of the industry.
3. “In 2025, more than 1,700 class lawsuits were settled for a total of $79 billion—nearly double the amount from the prior year.” That’s the click bait data point from Forbes’ review of the past year in class action litigation. The article goes on, though, to talk about some of the substantive issues for companies and industries arising from this class action boom. After over a decade of vigorous class action litigation over 401(k) plan fees and expenses, one issue that is often overlooked when the discussion just shifts to the amounts at issue is the extent to which that litigation campaign and history has reshaped so much of industry practice in this area. It’s the single best example of the private attorney general model of corporate regulation that we have ever seen, in my view. Now those same lawyers are coming for health and similar benefits. Buckle up.
4. In an article in Above the Law, lawyer turned tech journalist Stephen Embry discussed the open question of what the future, in the face of AI, looks like for lawyers in light of a new AI native law firm that, apparently, will basically do securities compliance work solely through AI tools. This gets us to the root problem when people talk about AI replacing lawyers or being able to practice law, which is that it depends on how you define lawyering. Taking ten rules of compliance and twenty examples of violations each, and then running trading or similar actions against them to determine whether compliance is met, is perfect for AI – it is simply applying a fixed body of rules to a set of circumstances. If that is your definition of lawyering, then certainly AI can do it. Likewise, if you define lawyering as just crafting slightly different versions of documents created for similar transactions a thousand times before, then certainly AI can project out the differentiation and spit out the document. But that represents the commodifiable portions of the practice of law, the types of almost rote training work assigned to associates and young partners to build their foundational knowledge and skills. Most experienced lawyers, if pressed to speak bluntly, wouldn’t even consider most of that to really constitute lawyering – they would instead consider lawyering to be the judgment calls, the advice, the new approaches to new problems, and the like that clients actually pay for (rather than the documentation, which just comes along with it). That’s the Rubicon that AI still has to cross before there can be any serious discussion of AI either practicing law or replacing lawyers, rather than just being another technological innovation that commodifies certain aspects of the practice, although admittedly in very dramatic and more extensive ways then we have seen in the past with other techological advances. (If the differential between lawyering and AI driven work product interests you, follow me on LinkedIn for more on this topic – I seem to be posting more and more on the question of how much of what AI does is really what we mean by lawyering, as opposed to just elements of the work of a lawyer).
5. I have written extensively on my views about nuclear verdicts, developed through decades of studying large verdicts for purposes of both advising insurers on how to resolve or avoid them as well as for purposes of litigating the bad faith and coverage actions they inevitably generate. Central to my views and writings on this subject is the rejection of the idea that nuclear verdicts are something unpredictable, attributable to the bad faith tactics of plaintiffs’ lawyers or to the biases of jurors, and are a surprise outcome in a given case. My view is that they are instead generated by predictable circumstances and fact patterns that reappear time and again across overly large and nuclear verdicts, meaning that insurers have the ability to avoid nuclear verdicts by learning to identify the claims that are likely to trigger them in time to settle before or during trial. After all, a too high settlement is still magnitudes less than a nuclear verdict. So I was extremely pleased to come across this detailed article explaining how nuclear verdicts are not arbitrary but instead attributable to predictable differences in the exploitation of information, and calling on insurers to match the plaintiffs’ bar in this regard. Nuclear verdicts aren’t a mystical thing – they are a predictable outcome of repeating patterns, and it is on insurers and their defense counsel to learn to identify (and settle) the risky cases before they can turn into nuclear verdicts.