Earlier this month, Reed Smith attended the 2026 AIRMIC Annual Conference at the ICC in Birmingham. The conference, under the theme Back to Basics, featured over 2,000 delegates from across the risk and insurance community who discussed the evolution of some established trends and explored emerging risks. We unpack some of the main themes and what they mean for policyholders and practitioners further below.
AI
AI was a dominant theme throughout the conference programme, with multiple sessions examining its implications for risk management, cyber exposure, and the insurance market. AI is already a major element in our analysis and decision-making process and organisations are positioning themselves strategically to navigate the next decade and the accelerated pace of development. The overarching message was one of urgency: the UK risks falling significantly behind in its response to AI-driven change.
The opening plenary on Monday afternoon featured Dr Keith Dear, CEO and founder of AI start-up Cassi, in conversation with AIRMIC CEO Diane Maxwell. Dr Dear, a former expert advisor to the Prime Minister on science, technology, space and modernising defence, and a former RAF intelligence officer, delivered stark warnings about the pace and geopolitical implications of AI. His central thesis was that society is still failing to grasp the scale of change underway and that AI is simultaneously "underhyped and underestimated".
Dr Dear framed AI as a major geopolitical force, exploring the emerging threats reshaping organisations and warning delegates that the risks are as significant as the opportunities. The session underscored how organisations need to be willing to incorporate the use of AI but also identified the human judgment, choices, and trade-offs that machines cannot replace.
Another key point noted by Dr Dear was that the US market is moving faster both in terms of AI adoption by insureds and in terms of the insurance market's response to AI-driven risks. For the UK, the key takeaways are threefold. First, the insurance sector must modernise its own processes and embrace AI-enabled underwriting, claims, and risk selection at pace. Second, policy wordings need urgent review to ensure they are fit for purpose in an AI-enabled world. Third, UK risk professionals must actively develop AI literacy and data strategies, working closely with IT teams and vendors, because not using AI at all is itself a key risk.
As the conference made clear, the question is no longer whether AI will transform risk and insurance – it already has. The question for the UK is whether it can close the AI gap before it becomes a competitive disadvantage.
Data centres
Reed Smith’s Peter Hardy joined experts from insurer FM, Caytons (a Gallagher Basset company) and Global Risk Consultants, exploring some of the insurance and risk issues emerging out of data centres. The headline message was stark: the sheer scale of the risk is unlike anything the market has previously encountered. Given their size and the concentration of value they house, if anything fails then the exposure is potentially huge. The panel considered the nature of the insurance response and level of protection available for data centre risk and considered how data centre operation and coverage may develop going forwards.
The challenges begin before a single cable is laid. Site selection (with power and water being absolutely critical factors) and early-stage design decisions can bake in the biggest risks from day one. From the perspective of an insured, meaningful mitigation of risk has to start before spades are in the ground.
The panel discussed how data centres are grappling with familiar insurance concepts (such as aggregation, mitigation, and valuation) but in a completely new risk landscape. For instance, how do you aggregate losses when an event impacts multiple data centres clustered in the same small area? How do you identify the replacement value of lost data to put an insured back in the position they were in before, where that may be practically impossible? How can debates around betterment be pre-empted before they arise?
Coverage gaps were raised amongst the panel members as a live concern. It is not uncommon for a property policy to exclude cyber risk and a cyber policy to exclude property damage, meaning a data centre risk sitting squarely between the two may fall through the cracks. In response, certain brokers are developing bespoke "super policies" designed to address the complete spectrum of data centre exposures. Terrorism cover is another frontier: data centres are increasingly recognised as critical national infrastructure, making them potential targets for bad actors. Some operators are now seeking terrorism coverage at limits in the billions, reflecting just how high the stakes have become.
The market has an appetite to insure these risks, but nothing tests coverage quite like claims experience and that has yet to come in any significant volume. The panel agreed that the market is currently very solutions-minded, but insurers need to be kept up to date on the projects, the construction, the technology and all the associated new risks if they are to maintain confidence in the cover they are writing.
Cyber
A foundational session on Day 1, The Growing Role of AI in Cyber Risk, examined how artificial intelligence is shaping cyber risk exposure and whether cyber policies need to adapt to the evolving exposure landscape. This theme was reinforced by pre-conference commentary from AIRMIC, which highlighted the increasing sophistication of cyber-criminals, including potential use of AI and large language models to enhance phishing and other attack methods.
Another conference session hosted by AIG, Cyber Risk Management: A Boardroom Checklist in Today's World, focused on the practical challenges of managing cyber risk in an environment where organisations are captured under multiple legal and regulatory frameworks concurrently. The session explored how boards and senior management are increasingly being held accountable to new standards and obligations, with particular emphasis on the interconnection between pre-incident preparedness and post-incident regulatory scrutiny.
There was much discussion at the conference about numerous recent cyber incidents and how, despite many pay-outs in the market, cyber (along with a number of other sectors) remains a soft market and there are opportunities to develop bespoke cover. In short, cyber remains a key topic of conversation and one of the most important insurance policies that corporates can have.
The insurance protection gap
AIRMIC and the International Underwriting Association spoke on the topic of insurance protection gaps, building on a series of roundtables held with AIRMIC members and insurance market representatives in early 2026.
Coverage gaps remain a challenge and the property, business interruption and construction insurance sectors were cited as examples where there are known gaps. Lack of knowledge about the availability of certain products was recognised, but there remained a fundamental question: whether that product would cover what an organisation needs or leave more gaps than it closes. In some cases, insurance will also not provide the answer; insurance is only one part of a business’ overall risk mitigation strategy, and some risks will be too interconnected for traditional insurance products to address.
For the majority of risks, the discussion favoured innovation within existing covers, such as add-ons or extensions, to address specific business risks. The example cited was cyber policy wordings, which started developing around a decade ago to address data loss. As claims data developed, underwriters were able to write more and expand coverage. The phrase ‘data-led underwriting’ was used to capture the approach of using claims and operational data to shape and price coverage.
That said, new products are going to be needed to address certain emerging risks, including AI (for instance, to address exposures arising from hallucinations and IP infringement) and data centres. It was recognised that the pace of change, driven by geopolitical developments and rapid deployment of technology, means that the market cannot afford to take years to respond. The market is under pressure to develop new products, as well as find the capital to service the required limits, to meet these emerging risk categories.
Climate change
The climate session, Accelerating Climate Risks, moderated by Max Neri, Climate & Resilience Risk Global Lead at Moody's, formed part of a wider programme addressing interconnected risk themes. The session addressed the proposition that climate risk consistently ranks amongst the most pressing risks facing global organisations. The panel explored how these risks manifest in two principal categories:
Physical risks: both acute and chronic including large-scale wildfires, flooding, heatwaves, and droughts. These perils are intensifying in frequency and severity, creating coverage gaps, and putting pressure on traditional insurance capacity.
Transition risks: encompassing the legal, financial, and reputational risks associated with energy transition, changing economic models, and evolving regulation.
The central question put to the panel was how AIRMIC’s risk managers should be preparing for and adapting to this accelerating landscape. With an acknowledgement that climate change is already upon us, there was an increased emphasis on physical loss prevention. Corporate risk managers are increasingly faced with complex coverage gaps, particularly around non-damage business interruption, and hyper-local weather perils, as climate volatility intensifies and traditional insurance capacity tightens. This involves stress-testing business operations against a range of climate scenarios (e.g., present day, 1.5°C, 2°C, and 4°C warming pathways) to identify vulnerabilities across assets, operations, and supply chains. A thorough risk mapping exercise should identify which business units, geographies, and assets are most exposed to physical and transition risks and those risks should be quantified using both historical loss data and predictive climate models.
It was made clear that climate risk should not be siloed and must instead be embedded within the broader ERM framework, with clear governance, reporting lines, and board-level oversight. Insurers should be brought into all of this work to ensure insurance programmes adequately cover corporates and particular attention should be paid to policy exclusions (e.g., gradual environmental damage, certain flood perils) and sub-limits that may leave significant residual exposure.
Climate volatility and its implications for insurance are clear but the insurance industry is still figuring out how best to support its policyholders who face difficult issues that require a decent (and sometimes reluctant) spend from the c-suite.
Claims
A recurring message at this year's conference was that while the claims landscape may appear familiar, its complexity has grown considerably. A single event can cause a range of different losses to an insured and trigger responses across multiple policies – the pandemic demonstrated this vividly, with business interruption, financial lines and insolvency-related claims all flowing from the same underlying event or cause in varying degrees.
The longevity of claims was a notable theme. COVID-19 cases remain live and geopolitical tensions in Russia/Ukraine and the Middle East continue to generate related claims. Misselling claims are also expected to continue to be a recurring feature of the claims landscape. Overall, the general trend is towards more active and heavily contested litigation. The interconnectivity of modern risk means that businesses must think carefully about their full policy suite and how different covers interact.
The claims experience has come under great focus in recent years, with Lloyd’s of London implementing standards for claims handling. The messaging for policyholders remains to engage early with internal stakeholders across the organisation, instruct appropriate and experienced advisers where necessary, and set the narrative for the claim at the outset.
Whilst AI may become a feature of the claims process in the future, it is not yet utilised in specialised and complex claims where human judgment remains essential. Understanding the limitations of AI and technology and ensuring a proper decision-making process sits behind any use (in both claims handling and to mitigate risk generally) were emphasised as key risk management priorities.