The term "AI Assurance" is used in many ways to mean different things, often leading to confusion and miscommunication. This session will provide a brief introduction to the topic, followed by a panel discussion with key players discussing their practical experience of providing and buying assurance, including their understanding of assurance, scope of work, limitations, and expectation gaps.

The potential for AI to improve effeciency and accelerate business growth is well recognised. However, adoption has been slowed down by concerns around the risks presented by AI including lack of transparency and explainability of models, bias, and accountability. This session will discuss to what extent current assurance mechanisms can mitigate these risks, including hearing from two businesses who have used AI Assurance to accelerate their organisation's responsible adoption.

What role does regulation play in promoting AI Assurance and responsible AI adoption? This session will debate the pros and cons of regulation as a driver for demand of assurance services, considering the environments in the EU, UK and US, with contributors arguing both for and against regulation, and participants voting on the best option at the end.

The AI Assurance industry presents new business opportunities, with the UK Department for Science Innovation and Technology has projected that the UK assurance market, could move beyond £6.53bn by 2035. Accounting firms are already becoming key players in this area. This session will discuss considerations for accountants and businesses looking to expand their services into this new area, with recommendations on how to get started

The EU AI Act requires conformity assessments to be performed for AI systems classified as High Risk AI Systems (HRAIS). This session will look at the requirements of assessments and lessons learned so far in preparing for these assessments

More and more audit clients are beginning to incorporate Generative AI into internal processes, including in financial reporting. The complexity of AI can present challenges to the way auditors have traditionally conducted audits, with areas such as repeatability being an issue. This session will explore the considerations for auditing a client using generative AI, and the types of assurance that may be considered as part of the process.

Third party models underpin the majority of AI tools used by businesses. However, getting assurance over such models can present challenges including not having access to required information. This session will explore these challenges, and ways to address them, including looking at the work of the UK AI Safety Institute, and exploring the potential to leverage third party reports

With so many different types of AI assurance activties, what are the vital skills for AI Assurance? How can the UK build such skills and what should businesse be doing? This panel will discuss these topics, as well as the practical considerations for the development of an AI Assurance profession.