The AI age brings promise and trouble
As senior vice-president of business transformation at Renaissance Re, Richard Prior sometimes uses artificial intelligence programmes to summarise the many reports he reads daily.
It saves him time.
When he has a question, he plugs it into the AI programme Chat GPT.
“I might go into Chat GPT and say give me a list of companies that I can talk to,” he said on Tuesday in a panel called Powered by AI: Exploring the Future of Artificial Intelligence in Insurance at the Bermuda Captive Conference 2023.
Panel moderator Michael Branco, of Celeste Ventures, questioned the safety of plugging company data into AI.
“What happens when all of a sudden, that data has reached the boundaries of security and data privacy for that company,” he said.
Appleby partner Duncan Card said people were still largely unaware of just how meaningful the impact of AI would be.
“Speech-activated data programmes have a database that records the questions asked, and the information given to it by the user,” said Mr Card the author of Information Technology Transactions: Business, Management and Legal Strategies.
“There are a lot of employees of multinational companies who are using these systems. They are talking to them and asking them to do stuff. A lot of very confidential business information is going from the corporations at that grass-roots level into these databases.”
He said people who worked as engineers, doctors, nurses, lawyers or assembly line workers were also using AI.
“They are asking for its help in doing stuff,” Mr Card said. “They are making decisions based on technology that they know nothing about. They do not really know who is behind it.”
He said some firms were now developing focused policies on how employees access and utilise AI programmes.
Justin Willmott, vice-president of SiriusPoint Re, said the world’s companies, governments and societies were not very well designed to handle exponential change.
“There are all sorts of unforeseen consequences to the rapid adoption of AI within the insurance world,” he said.
On the plus-side AI can now carry out insurance tasks that would once have taken hundreds of manpower hours to complete, in a matter of minutes.
Mr Willmott is seeing clients using artificial intelligence to develop new products such as voice to text apps.
However, he is also seeing it being used for more nefarious purposes.
“It can be used for massive scale social engineering,” Mr Willmott said. “For example, it can be used to deploy phishing emails at a scale, that before now, was just unimaginable.”
He said that the explosion of growth in AI magnified cyber-risk.
“There are all sorts of unforeseen consequences of the rapid adoption of AI within the insurance world and those types of things that insurance companies indemnify,” Mr Willmott said.
“If you’re doing a deal and you’re getting a bunch of representations and warranties, there are different ways that AI events could manifest to impact them.”
Mr Willmott thought that we might soon start to see AI exclusions included in insurance policies.