AI Assurance Forum explores practical approaches to third-party AI assurance

Leaders from business, learned societies, academics, and government gathered to design practical approaches to third-party AI assurance services.

Leaders from business, learned societies, academics, and government gathered to design practical approaches to third-party AI assurance services. With co-organisers from the ICAEW, the Office for Statistics Regulation, DSIT, and Validate AI, Imperial hosted an event with the immodest goal of sketching a blueprint for practical AI assurance services that incorporate lessons from centuries of London-based innovation in assurance professions.

“As AI systems become embedded across organisations and public services, questions of governance, accountability and trust are increasingly important.” - Professor Mark Kennedy, Director at Data Science Imperial

Professor Mark Kennedy, Director of Data Science Imperial and host of the event, said:

“As AI systems become embedded across organisations and public services, questions of governance, accountability and trust are increasingly important. This event gathered leaders in this area to collaborate on how to make the most of AI by designing services for assessing whether AI systems are deployed competently and responsibly—in the spirit of the work of assuring the reliability of financial statements. As in finance, the stakes here are massive: this is about supporting responsible growth in AI-related services and markets.”

The event was built around a series of design “sprints” based on using AI systems in scenarios spanning healthcare, mental health services, taxation, and insurance. These exercises were designed and facilitated by Professor Ileana Stigliana, head of Imperial’s Strategic Design Studio, who said, “To responsibly embrace the growth and innovation AI can bring, we need not just world-class technical skills, but also first-class expertise in designing services that, in this case, are likely to become an important professional service.”

The event reflects Imperial’s continued effort to foster collaboration across disciplines and sectors to ensure the future of artificial intelligence is safe and prosperous.