Artificial Intelligence (AI) and machine learning systems have recently been shown to be susceptible to biased and discriminatory decision-making. Detecting and fixing these problems is difficult due to the inherent opacity of these systems. In response, calls to make algorithmic decision-making systems more transparent, explainable, and therefore accountable can be seen in academic literature, policy proposals and laws, such as the EU’s General Data Protection Regulation. At the same time, explaining the functionality of complex algorithmic decision-making systems and their rationale in specific cases is a technically challenging problem. Some explanations may offer little meaningful information to individuals, raising questions around their value. Further, companies and governments have an interest to not disclose information about their algorithms that contain trade secrets, violates the rights and freedoms of others (e.g. privacy), or allows individuals to game or manipulate decision-making.
This talk explains why accountability in AI-based systems is an ethical responsibility, explores the legal requirements in the General Data Protection Regulation applicable to AI and examines if ‘counterfactual explanations’ can bridge the gap between transparency and competing interests.