There are few AI insurance policies that are commercially available and losses are not explicitly covered by traditional insurance coverage 

Insurer Zurich has called for increased cross industry collaboration around artificial intelligence (AI), to raise public confidence in the industry mitigating possible risks, as well as better understanding how to tap into potential opportunities. 

As society becomes increasingly digitised, emerging risks around decisions made by algorithms and AI are becoming more common. 

This is according to Zurich and Microsoft Corp’s white paper, titled Artificial Intelligence and Algorithmic Liability, published last week (29 July 2021), which takes an in-depth look at relevant cases in product liability, professional indemnity and medical malpractice that can provide guidance in minimising the exposure and potential harm to customers and organisations’ reputations arising from AI-related risks.

Ericson Chan, group chief executive and digital officer at Zurich Insurance Group, said: “Technology can improve the quality of our lives in many ways. Data and AI further accelerate change.

”We need to be diligent on managing algorithmic risk and [the] ethical challenge they bring by ensuring fairness and privacy, with transparency and clear accountability.” 

The report points out that unleashing the power of data and AI creates “endless business opportunities to ultimately improve the quality of our lives”.  

But, with those opportunities comes a “broad spectrum of risks encompassing not only regulatory compliance, but also liability and reputational risk if algorithmic decision-making triggers [lead to] unintended and potentially harmful consequences”.  

Few AI insurance policies 

From an algorithmic liability risks perspective, there are few AI insurance policies commercially available and AI losses are not explicitly covered by traditional insurance products. 

AI-related claims that could arise include allegations of negligent design - this is when the product is manufactured as intended, but the design causes a malfunction or creates an unreasonable risk of harm that could be have been reduced or avoided.  

In some cases, business interruption (BI) risks potentially related to unstable AI systems may be covered under existing BI policies.

Meanwhile, algorithmic risks vary across insurance lines - there are specific risk management services that insurers can potentially offer to large corporate and SME clients.    

For example, errors and omissions (E&O) insurance in the field of robotics offers professional liability coverage that could be considered for AI systems - this would typically be complemented with specialised risk management services.  

The report added that the use of AI systems which fail to function and perform within expected and acceptable business outcomes can lead to property damage, medical malpractice and professional liability claims.  

Problems can arise due to: 

  • Input data.  

  • Algorithmic design.  

  • Output-driven decisions.