Conference on the Rule of Law and Automatized Decision-Making, 24 August 2021
The use of artificial intelligence in the service of the public sector may assume various forms, but a relatively well circumscribed area is automatized decision-making, that is, the making of administrative or even judicial decisions by means of algorithmic decision-making. Such decision-making can broadly speaking be either rule-based or based on machine learning, and it can be either fully automated or partially automated. One clear precondition for the use of algorithms as the basis of decision-making is the same as with human decision-making at public authorities: it must be possible to connect the decision-making to the rule of law. The legal orders of various countries contain different approaches to the relationship between the rule of law and automatized decision-making.
It appears that the variation displayed by several countries may range from relatively liberal and open systems allowing a flexible use of automatized decision-making to systems where the expectations concerning procedural and material legislation of an explicit kind are high. The rule of law along with other core constitutional values, principles and rules such as democracy, individual rights and freedoms, right to effective remedy and fair trial form the basis of our societies.
The Council of Europe is at this moment in the process of working out its own approach to AI-related issues where these values, principles and rules will probably be featured. At the same time, automatized decision-making is already taking place in the administrations of various countries. It is high time to intensify the discussion about the fundamentals of society that is increasingly using algorithmic decision-making for its functioning. At the same time, the individual whose matters are being managed by means of automatized decision-making should be the main driver of the need to discuss automatized decision-making.
Recently, the European Union released a proposal for an overall harmonizing regulation of artificial intelligence, ostensibly also covering applications within the public sector, although private sector uses of AI are probably more important from the perspective of the business life. The rule of law does not appear to be at the centre of the proposal, but the point is made that certain AI systems “intended for the administration of justice and democratic processes should be classified as high-risk, considering their potentially significant impact on democracy, rule of law, individual freedoms as well as the right to an effective remedy and to a fair trial”, as elaborated in Annex III of the proposal. Delivery of public services by means of algorithmic decision-making is thus an important dimension of the proposal, although the term automatized decision-making is not used. If adopted, the proposal might offer a rule of law basis for automatized decision-making also within the public sector.