Industry urged to understand AI

By Resolve Editor Kate Tilley

Will artificial intelligence (AI) take our jobs and make us redundant?

Bernie Heinze, Executive Director of the US-based Federation of Defence and Corporate Counsel, posed that rhetorical question at WICA2023, and the answer was yes.

“But that’s not meant to scare you. It’s an honest appraisal of the direction in which AI is heading, but not necessarily something to be afraid of,” he said.

While we don’t understand AI and “don’t quite know what to do with it”, we need to learn about it.

Mr Heinze said the insurance industry had adapted to many things over the years and AI was another, impacting particularly in underwriting and claims.

When the AI-powered generative language model ChatGPT was introduced, it had more than 100 million unique users within the first 30 days. He said AI was yet another disruptor and that was at the industry’s core. “Without disruption, there would be no need for insurance.”

Data explosion

Mr Heinze predicted an explosion of data from connected smart devices like phones, watches and vehicles, which will give insurers better opportunities to understand their customers and therefore provide better products and pricing, and real-time service delivery and claims experiences.

Another trend was an increasing prevalence of physical robotics. He predicted 3D-manufactured buildings that use stronger materials that can withstand storms and fires, autonomous farming and more parametric insurance.

He also predicted an increase in risk purchasing and risk retention groups using alternatives to conventional insurance to better control risks, underwriting and claims.

Mr Heinze said generative AI like ChatGPT was continuously learning from answers given to questions being asked. But there was no clear answer on protecting personal information and ways to ensure it was not being shared across various platforms.

He defined AI as a capacity given by humans to machines to memorise and learn from experience, to think, create, speak, judge and make decisions.

WICA2023 limerick

Mr Heinze tested Google Bard by asking it why he should attend WICA2023. The chatbot took 3.7 seconds to provide an answer that was “quite on point”. On request ChatGPT write a limerick about WICA2023 which it translated into five languages.

He said ChatGPT launched in November 2022, which meant answers were limited to information available then, but developer OpenAI was now finding ways to make it more current.

Mr Heinze said AI enabled machines to actually think, rather than just collect data and respond back on a search across varied open sources.

“Machines can have consciousness, empathy and embrace situations, using natural language processing to understand everyday speech.”

The World Economic Forum has said 48% of insurance functionality is highly adaptable to AI.

Positive decisions

Mr Heinze said AI would enable a greater focus on prediction and prevention. “It’s going to enhance our ability to make good positive decisions and give good advice. It’s going to lower costs and empower our ability to price more specifically and optimise the customer experience.”

Mr Heinze predicted a blurring of the lines between traditional underwriting, broking and risk management. “That’s all going to change as we look for ways to use AI to improve the customer experience.”

Usage-based insurance will increase so customers are “not paying for insurance with the car sitting in the garage or a parking lot”.

Mr Heinze predicted that, by 2030, more than 50% of claim activities will be automated. Advanced algorithms will handle the initial claim functions with an increase in speed and efficiency and a decrease in costs and waiting times.

There will be greater consistency and the ability for customers to see their claim status in real time.

Better accountability

“AI will be able to draft responsive pleadings, motions, briefs, summary judgement motions, and discovery and hopefully create lower loss ratios. Will mistakes still be made? Probably yes. But, with each transaction, with each way the machine learns, it gets better and thereby will drive better accountability. We’ll also have scams, hackers and malware.”

Mr Heinze said the American Bar Association had established a committee to work with lawyers to understand what AI is, how it’s going to be used in the legal profession, and how it can be used ethically, honourably and transparently.

“We need to retain the personal side of underwriting and claim functions. AI cannot do it all. We cannot allow quality to succumb to expediency.

“We have to have ethical and bias issues addressed to ensure AI is honourable. We have to get ahead of the implementation now by seeing how we can use it best.”

Mr Heinze called for trusted collaboration to ensure AI was implemented ethically, honourably, transparently and professionally.