EU proposals for AI lay bare the digital conundrum

Europe can’t afford to miss out on an era-defining technology…

In a previous blog, Build a trustworthy framework, and they will come (January 2019), I observed that the European approach to Artificial Intelligence was based on the idea that ‘trustworthy AI’ would be a spur and not a barrier to innovation. A doctrine of ‘responsible competitiveness’ would make the EU an attractive place to invest. This is because consumers would be confident in AI regulation, and more willing to adopt the technology. The question was whether industry would agree, or be tempted to invest in more unconstrained markets.

From the proposals published last month (‘Proposal for a Regulation laying down harmonised rules on artificial intelligence’), we have more of a clue about which way the commission is leaning. Most AI uses will fall outside of any regulation, and even ‘high-risk uses’, such as the use of AI in public by law enforcement. Aimed squarely at facial recognition technology – they are subject to exemptions under circumstances that many campaign groups consider too broad.

Rules to address the risks of discrimination by AI systems are similarly criticised for not going far enough.

Other campaigning organisations have suggested that there is too much reliance on inconsistent national oversight bodies. Furthermore, not enough emphasis on giving consumers enforceable rights and opportunities for redress. Inevitably, technology lobby groups have pointed to the dangers of red tape, (which the Center for Data Innovation colourfully claimed would ‘kneecap the EU’s nascent AI industry before it could walk’).

More moderate voices have questioned whether the proposals offer the right balance of risk and benefit.

However, the concerns of the European Commission are most evident from the comments of Thierry Breton, the Internal Market commissioner: ‘So now, if you want to use artificial intelligence applications, go to Europe! You will know what to do, you will know how to do it…you will also come into a continent with the largest amount of industrial data created on the planet for the next ten years’.

In the post pandemic, digitally-driven future, attracting AI investment is crucial, and protection expectations need to be tempered with economic reality.

With GDPR, the EU had some success in placing data protection standards on the global agenda.

The hope is clearly that the same will happen in AI standards, and that the successful digital platforms of the future will come from Europe, and not just Asia or the US. The proposal will need to be adopted by the European Parliament and EU member states. Historically, this can take a long time.

Depending on your point of view this could be a good thing, allowing for a greater consideration of likely consequences, or it may ‘cost’ Europe the potential to be the dominant voice in AI that the commission hopes. In this respect, the AI regulation is probably the most important – and high stake – piece of digital legislation the EU will adopt in the next ten years.