The EU Council has adopted its position on what is known as the EU AI Act - a set of harmonised legal rules ensuring that AI systems placed on or put into service in the EU market 'are safe and respect existing law on fundamental rights and Union values.'

The next step is for the EU Parliament to review and adopt its position on the proposals, and then for the EU Council and EU Parliament to enter negotiations. These will be in light of - and in parallel with - the EU's proposals for an AI Liability Directive and updated Product Liability Directive which are to complement the AI Act by providing a holistic and comprehensive legal framework to address the opportunities, risks and potential harms of AI systems.

The updated AI Act text includes a number of amendments to the original version which saw thousands of proposed amendments (as we wrote about here from various EU Committees and bodies such as: Legal Affairs; Industry, Research and Energy; and Culture and Education; the European Central Bank).

Those updates include:

  • Refining the definition of AI - 'the Council’s text narrows down the definition to systems developed through machine learning approaches and logic- and knowledge-based approaches.'
  • Updating the classification of high-risk AI systems - such systems are subject to specific requirements under the AI Act. The amended text seeks 'to ensure that AI systems that are not likely to cause serious fundamental rights violations or other significant risks are not captured'.
  • Making requirements for high-risk AI systems more technically feasible - 'for example as regards the quality of data, or in relation to the technical documentation that should be drawn up by SMEs to demonstrate that their high-risk AI systems comply with the requirements.'
  • Clarifying who is responsible for what and the interplay with other EU laws - 'Since AI systems are developed and distributed through complex value chains, the text includes changes clarifying the allocation of responsibilities and roles of the various actors in those chains, in particular providers and users of AI systems. It also clarifies the relationship between responsibilities under the AI Act and responsibilities that already exist under other legislation.
  • 'Clarifications and simplifications' to conformity assessments

Will the EU AI Act work in practice?  The proof will be in the pudding.  But there are many analyses and reports - like this one - seeking to test how industry view the proposals.  And it is useful to keep up to date with how the debates and negotiations continue, to proactively identify potential gaps and issues, and overall direction of travel for how regulations should address the opportunities and harms of AI.  

If you would like to discuss how current or future regulations impact what you do with AI, please contact Tom Whittaker or Brian Wong