Following the 29th of November 2023, approval of the revised definition of an ‘AI System’, the Organisation for Economic Co-operation and Development’s (“OECD”) on the 5th of March 2024, published its ‘Explanatory Memorandum on the Updated OECD Definition of an AI System’.
How did the definition change?
Originally modelled on Russell and Norvig’s conceptual outlook on AI,[1] the previously established was updated, in contrast with the previous definition, it now reads as follows:
“An AI system is a machine-based system that can, for a given set of human-defined explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as makes predictions, content, recommendations, or decisions that can influencinge physical real or virtual environments. Different AI systems are designed to operate with varying in their levels of autonomy and adaptiveness after deployment.”
The revised interpretation of an AI system has been designed to be all-encompassing, incorporating a range of different systems. Notably, the OECD stated that certain methods like Machine Learning necessitate nuanced considerations for policymakers, especially in order to tackle pressing issues of bias, transparency, and explainability.
Revisiting the AI Act’s definition
Conversely, the EU’s Artificial Intelligence Act (“AI Act”), text of the provisional agreement, defines AI systems as the following:
‘AI system‘ is a machine-based system designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments; (Article 3(1))
The primary distinction between the two definitions lies in the emphasis on autonomy and adaptiveness. The AI Act’s definition underscores the varying levels of autonomy and the potential for adaptiveness post-deployment as inherent attributes of AI systems. Most notably, it highlights the machine's capacity to operate independently and adapt to changing circumstances.
In contrast, the OECD’s updated definition maintains a broader perspective, acknowledging that AI systems can differ in their levels of autonomy and adaptiveness without explicitly detailing these variations. The recent changes prioritise the fundamental function of AI systems in generating outputs based on inputs to influence environments while acknowledging that autonomy and adaptiveness can vary among different implementations.
Historically, the OECD’s definition has been favoured by EU legislators over other definitions, will this updated definition find its place within the AI Act before its enactment?
--
The published OECD explanatory memorandum can be accessed here.
Need assistance relating to Artificial Intelligence Law? GTG is here to help! Contact Dr Ian Gauci for further information.
[1] Russell S and Norvig P, Artificial Intelligence: A Modern Approach, 3rd Edition (Pearson Education 2009)