We have a European AI rulebook. What do companies do now?
After three years of lively debate on the AI Act, today’s vote of the European Parliament will usher in a new era of European digital innovation.
We are pleased to have one set of rules and not 27, but it’s no secret that the AI Act will add extra burdens that companies outside of Europe won’t have.
I am particularly concerned about the growing AI healthcare sector in Europe, already heavily regulated. AI in healthcare will fit into the high-risk category. Companies like Corti, a Danish startup that uses AI-based voice recognition to help medical professionals make split-second life-saving decisions, or the Hungarian start-up OnCompass, which uses AI to spot cancer in scans, are already expanding in the US. How can we keep these jewels from leaving Europe?
It is therefore vital that we think deeply about how to keep our brightest companies here in Europe and take action to create even more of them. Today, only 3% of the world’s AI unicorns come from the EU, with about 14 times more private investment in AI in the US and 5 times more in China.
By 2030, the global AI market is expected to reach $1.5 trillion, and we need to ensure that European companies tap into that without getting tangled up in red tape.
Here are five things the EU should keep in mind for the weeks and months to come:
1. Make use of what’s already out there
European Harmonised standards are crucial to make compliance as easy and predictable as possible. They define exactly how a company can fit the AI Act requirements. The problem is, they don’t exist yet. Instead of starting from scratch, there are over 100 global tech standards -75 already supporting EU AI Policy – available. Let’s leverage them to get European standards as fast as possible!
2. Make compliance easy
For example, why not create an online portal for self-assessment, so companies can get a quick answer on which risk category they fall into? In all EU languages, and simple to use.
We also need to increase funding for sandboxes. They will need enough capacity to handle all companies’ requests, wherever the company is located. Additionally, public authorities should respond to requests within 3 months, so that businesses can plan their compliance activities with more certainty.
3. Lead the world to speak the same AI language
AI is global by default. Critical technologies like AI are also key for our economic security – we need to align with like-minded countries. In the past year, we have seen a plethora of AI rule and guidelines-making initiatives popping up across the globe from the US to Japan to China, as well as international bodies like the G7, the UN, and the African Union. Together, we need to find common answers to questions like how we define AI and how we categorize and manage the data that fuels it.
4. Boost investments in AI and align on tax incentives
If we want to double the amount of unicorns in the EU, investment in AI is vital. Last year, only 6% of AI venture capital went to European startups. The next Commission needs to set the framework for greater private investment in emerging technologies – for example by completing the capital markets union. Tax incentives for innovation also need to be part of the solution, but we need to be sure they are better coordinated between member states.
5. Detangle the regulatory spaghetti bowl
The AI Act doesn’t come alone. Companies will need to navigate up to a dozen pieces of legislation, from the GDPR to the Data Act, but also sectoral rules on medical devices or machinery. The European Commission and Member States should team up to sort out any overlaps or clashes between these laws and the AI Act.
“Europe has the potential to be an AI powerhouse, but the AI Act alone will not deliver it. The next few years need to be laser-focused on creating the conditions for Europeans to be the creators of AI, not just the users.”
Cecilia Bonefeld-Dahl, Director General DIGITALEUROPE
