European AI Act
The proposal will be discussed in the Parliament and Council before becoming law. The new rules will focus on making AI trustworthy while still promoting AI uptake investment and innovation in a coordinated member state approach. A very tall order. Here, I have put together a summary of the new Act and will share ongoing discussion as to its impact on industry.
People first approach
The EU will continue its human-centric position on AI with this new legislation by following a risk-based approach. Any AI system that is considered an unacceptable risk will be banned. An unacceptable risk will be anything that poses a threat to people’s safety or rights and could potentially manipulate human behavior. To mitigate public concerns about the use of unethical AI applications, the EC proposes to ban a number of AI applications that manipulate human behaviour or conduct social scoring.
- High-risk AI, such as AI used in education or law enforcement, will have strict obligations and restrictions before they can be put on the market. The use of biometric identification will still be allowed in a law enforcement capacity. Businesses will have to show that they comply with EU standards through self-assessments and national checks.
- Limited-risk AI will have specific obligations depending on its uses. Chatbots, for example, will have transparency-based obligations, people will have to be told they are talking to a robot.
- Minimal-risk AI is the category most AI falls under and will not be affected by the new rules. They will continue to allow free use of minimal-risk AI.
Investment and innovation
The proposal also makes sure to leave space for innovation in AI. Regulatory ‘sandboxes’ will allow smaller companies to responsibly innovate AI without consequences. An updated Coordinated Plan on AI will also contribute to innovation and investment in AI. The new plan will use funding from Digital Europe and Horizon Europe programmes, as well as the Recovery and Resilience Facility to help promote innovation in AI.
This is really important for the future growth of SMEs across Europe. We need an environment conducive to developing new products and services with AI that is also supported by the right funding mechanisms.
What happens next?
The proposal will now be debated in the EU parliament and council before becoming a law. It will be interesting to see how the EU will balance safety legislation with innovation in AI and the shape these “regulatory sandboxes” will take.
As part of our work with the DIGITAL SME working group, we will continue to share developments on the proposal and what it means for SMEs.