Briefing: Brokers are feeling the AI pain as questions come thick and fast

The benefits and the risks of AI continue to be one of the biggest topics of concern for businesses and their insurers.

The Biba annual conference in May saw much discussion around its use in the industry, with focus on the most efficient way to get the best out of the technology as development rapidly evolves.

In the past week, broker Willis published its latest Risk and Resilience report, in which it noted that the world has reached a point where it needs to ask what artificial intelligence can change or improve – rather than simply asking what it does.

The broker’s review highlighted how AI is being embedded across underwriting, claims, cyber defence and operational decision-making.

And Willis added that this growing integration into the sector was introducing new challenges around accountability, liability and insurability. Unsurprisingly, it concluded that the challenge now facing businesses is the responsible adoption of AI.

Brokers have been quick to seize on these opportunities and one specialist broker at Marsh even told me that the arrival of AI had changed the world.

“AI is, in my view, the largest revolution since the industrial revolution, in terms of its impact on society and the economy,” he said.

Customer impacts

The rise of AI may well herald a new dawn for mankind, but the scale of the global and major industrial risk is considerable.

For example, with the energy and data-intensive tech driving the construction of £20bn plus data centres across the globe, there is now a huge, new demand for water and energy. Indeed, a single hyperscale data centre can consume millions of gallons of water daily

However, while AI is seen to be a global issue, its impact is being felt closer to home – with unintended consequences leaving many brokers frustrated.

The technology has quickly embedded itself in many individuals’ lives, especially with large language models (LLM) now commonplace, with the likes of Copilot, ChatGPT and Claude available on every major operating system.

LLMs can be incredibly useful for drafting emails, writing code, creating blogs or brainstorming, but are also very adept at extracting key points from lengthy documents, books, or datasets.

And brokers are now finding that clients are turning to their LLMs to ask: “What should I be asking my insurance broker?”

Some of the answers generated to this question are leaving intermediaries scratching their heads and having to respond to a range of strange, and often confused, questions.

One leading reginal broker bemoaned: “The past six months has seen a surge in strange questions being submitted by current and perspective clients which are clearly AI generated.

”Far from making the process easier and more efficient, we are now being bogged down with questions that leave us asking what we are supposed to say in reply.

“As brokers we always welcome questions from our clients around their risks and the solutions we can deliver but AI is now clouding the process.”

Responsible use of AI may well be the biggest challenge for the insurance sector and its clients – and there are likely some testing times ahead.