The European Union (EU) is in the process of establishing important guidelines around the use of artificial intelligence (AI) algorithms in the EU single market. While there are considerable benefits to incorporating AI into various automated decision-making systems, there have been multiple high profile cases where AI-based decisions adversely affected individuals or groups. This means that trust in these systems is critical for their widespread adoption in society. With their recently proposed AI Act, the EU is attempting to strike the right balance between promoting technological innovation and protecting the EU’s values, fundamental rights and principles. This talk presents a brief overview of the proposed legislation and then focuses on one particular aspect of the draft bill; namely bias monitoring. It is stated within the act that “providers should be able to process special categories of personal data, as a matter of substantial public interest, in order to ensure the bias monitoring, detection and correction in relation to high-risk AI systems”. AI areas considered to be high-risk include medical devices, financial services, education, employment, law enforcement as well as critical infrastructure (transportation, water, gas, electricity). This talk briefly introduces some methods such as correlations, confusion matrices and statistical modelling aimed at helping such providers in this regard, with the aid of some real-world data examples which will hopefully provide the stimulus for a rich discussion to conclude this talk.
