As the artificial intelligence (AI) boom expands new insurable asset classes and increases exposure to new liability and cyber risks, Swiss Re Institute, part of the global reinsurer, expects AI-driven disruption to reallocate rather than add demand for insurance.
New research from Swiss Re Institute explores the rise of AI and its influence on macroeconomic and insurance risks, stating that the US is economy is exposed to large, concentrated investment and asset-price effects, while productivity and output gains from the technology emerge at a fairly slow pace.
According to Swiss Re Institute, a sustained 20% decline in the heavily AI-driven S&P 500’s valuation multiple could erode at least $10 trillion of household net worth, and cut US aggregate demand by 1–2 percentage points
(ppts) or more in the future.
On the insurance sector specifically, the research states that, “AI is expanding new insurable asset classes, from trillions of dollars invested in data centres to millions of kilometres of new power lines. AI is also increasing insurance exposure to new liability and cyber risks, and less than 5% of insurers we analysed disclosed a financial impact from AI adoption. Over time, we see AI-driven disruption reallocating rather than adding demand for insurance, likely requiring active portfolio and underwriting adjustment.”
Expanding on the first point, Swiss Re Institute draws on McKinsey research which suggests that under certain scenarios, data centre infrastructure investments could reach $7 trillion by 2030, suggesting that in the near term, insurers will likely see growth in premiums from new asset classes.
“This rapid expansion is likely to increase demand for insurance in the near term. We expect corporates to seek to cover both new infrastructure and the increased risk exposures of AI systems,” continues the report, which was authored by Mahir Rasheed, senior economist, Swiss Re Institute.
Swiss Re expects greater demand in engineering and construction insurance lines, while liability insurers can expect higher demand related to risks to third parties, with greater demand for trade credit insurance also likely.
“However, over the medium to long term, the demand uplift may begin to reallocate rather than simply grow. We expect AI to disrupt industries, creating “loser” industries (where AI reduces the need for physical assets, labour, and risk exposures) with reduced risk exposures and fewer assets to insure, and compressing traditional business models. This may shift medium-term demand patterns as some lines of business see volume decline, while new lines grow.
“The net effect may be a reallocation of premium volumes rather than continued strong growth. Insurers and reinsurers may wish to evaluate both this upside, and the reallocation risk: where new growth sits, where legacy demand may shrink, and how to reposition underwriting, capacity, and product design accordingly,” explains Swiss Re Institute.
The company’s analysis shows that currently, property and casualty insurers are leading the way in terms of AI adoption in underwriting and claims, while life and health use cases tend to focus more on distribution and operations for customer engagement, as well as chatbots and enterprise productivity.
“Insurers are seeking operational benefits from AI in the form of efficiency gains, time saved and workflow enhancements, yet less than 5% of insurers in our analysis of insurance use cases have disclosed any financial impact. At this stage, use cases are largely incremental but there is potential for wider enterprise transformation. In the near term, we do not expect AI-driven labour market dislocations, as most insurers aim for human workforce augmentation rather than complete automation of processes. On the contrary, AI may play a role in mitigating future talent shortages, depending on adoption pace and regulatory developments,” reads the report.
Importantly, Swiss Re Institute also discusses new risk exposures and challenges to carriers’ underwriting profitability as a result of AI. This includes things like lower barriers to legal system abuse in cases of liability litigation, fraud, data biases, cyber vulnerabilities, business interruption and concentration, performance issues
and intellectual property concerns.
“AI can amplify cyber threats through deepfakes, data manipulation, and fraud, while growing reliance on a few large cloud and AI service providers increases systemic and concentration risk,” says Swiss Re Institute. “Failures in algorithms or biased decision-making can result in D&O, E&O, and professional liability claims, as boards and professionals face scrutiny for governance lapses or flawed model design. Similarly, non-physical business interruption risks from AI or digital outages can propagate through supply chains, creating new forms of accumulation across insurance portfolios.”
As the report highlights, data on loss experience remains very limited, although will improve over time, and forward-looking models might well struggle to properly understand the shifting landscape.
Swiss Re Institute also highlights the potential for complex claims scenarios and potential accumulation risk, driven by AI risks rarely aligning with traditional policy boundaries.
“From a profitability perspective, AI’s efficiency gains and lower expense ratios may not necessarily translate into sustained underwriting profitability. Competitive market dynamics tend to pass cost savings on to policyholders, limiting margin improvements and leaving the industry operating close to its cost of capital. Additional potential lies in deploying AI to enhance risk selection, prevention, customer experience and claims management, thereby improving resilience in addition to efficiency. In those conditions, AI-led productivity gains may encourage competition and feed premiums growth, which has more ambiguous implications for underwriting profitability,” concludes the analysis.
