4 Artificial Intelligence and discrimination 4.1 What is AI? How is it used? While the term AI is widely used, it does not have a precise, universally accepted definition. This Guidance Resource uses the term AI in a consistent manner to the Human Rights and Technology Report to broadly refer to a cluster of technologies and techniques, which include some forms of automation, machine learning or algorithmic decision making.64 Insurers have long relied on data, statistical analysis, and models to determine risk and set prices for their policies, even prior to modern computing. Interpreted broadly, a wide range of technologies and techniques traditionally used by insurers in pricing or underwriting may be considered AI, or may otherwise form part of AI-informed decision making. In the insurance context, AI may be used in a wide range of different ways, including in relation to pricing, underwriting, marketing, customer service or internal operations. This Guidance Resource focuses on the use of AI in the context of pricing and underwriting decisions as these decisions are more likely both to use AI and to have a legal or similarly significant effect for an individual. Such decisions may also be more likely to give rise to discrimination complaints from customers. However, many of the general principles outlined in this Guidance Resource may also apply to the use of AI-informed decision making in other contexts. 4.2 Algorithmic bias While the use of data and models by insurers is not new, AI has the capacity to analyse large volumes of granular data more quickly, and to create more complex and potentially more accurate models. AI systems also include techniques not traditionally used in insurance ratemaking such as machine learning, including deep learning or neural networking processing. This means that while the risks may be heightened, they are not new. AI can enable good, data-driven decision making. It can be used to analyse large amounts of data quickly, accurately and cost-effectively. However, whilst AI promises faster and smarter decision making, AI-informed decision making carries with it certain risks. It can assist in identifying and addressing bias or prejudice that can be present in human decision making, but it can also perpetuate or entrench such problems. It can sometimes result in decisions that are unfair or even discriminatory. This is often referred to as the risk of algorithmic bias. This Guidance Resource refers to ‘AI-informed decision making’. This means a decision, or decision-making process, where AI is a material factor in the decision, and where the decision has a legal or similarly significant effect for an individual. The decision does not have to be wholly made by AI – the decision-making process could involve both AI and human involvement, but the involvement of AI must be material or significant. ‘Algorithmic bias’ does not have an agreed meaning but is usually understood to refer to the situation where AI is used to produce outputs that treat one group less favourably than another, without suitable justification.65 Algorithmic bias can include statistical bias and may result in unfairness and, in some circumstances, unlawful discrimination. It can arise through problems with the data being used by the AI system, or problems with the AI system itself. Guidance Resource: Artificial intelligence and discrimination in insurance pricing and underwriting • 2022 • 21