The European Take on AI
Can the EU Ecosystem still excel and bring value to the AI table?
A very interesting panel on AI took place at the IdeasLab convention, organized by the CEPS think tank in Brussels, Belgium today. The purpose was to have public politics experts, representatives of the political institutions, and leaders in private sector.
Startups, enterprises and research will need to cope with this visions and ecosystem of the years to come, both when implementing AI solutions and when applying or using it. That’s why it’s important to broaden the vision from a merely technical or economic exploitation of AI, to a broader perception of political and societal impact of AI implementation within day to day businesses.
The take of the Directorate General CONNECT of the European Commission
- it’s important to enable and support the European Industry to adopt AI as a required asset for actual competitiveness
- This will span the whole industrial fields in the next years, and we need to prepare to that
- Appropriate expert panels have been devised for structuring a vision towards this future
The take of the European Parliament on AI
- ethical principles: how can we add in the ethical level or at least some balance to the algorithmic content
- you must cope with different values and perceptions over behaviour and rules on privacy (quite strong in the EU wrt the rest of the world)
- additional protection layers for citizens are a typical asset of the EU wrt other countries. This may slow down evolution, but grant more safety
- ask for a special excellence institute on AI at EU level, to avoid scientists migrations to (non-EU) companies
- make sure that rules apply also to tentative of distortions of perception of reality.
The take of Google
- private sector also accepts and welcomes possible rules and principles
- last June they released publicly the Google AI principles and four red lights that Google will not cross (for instance, no development of AI for weapons), as well as Google responsible AI practices (also publicly available), updated quarterly.
- Based on this and possible threats, Google will not have public APIs for general purpose AI-driven tasks such as face recognition. Instead, it will deliver vertical, customized solutions that benefit specific cases or societal challenges
The take of Morgan Stanley
- it’s very important to have a single set of clear and objective rules on how to store, use and access data, and how to implement algorithms
- a shared research open sandbox should be available to the community for collaborative research
- once AI comes into play, the whole problem of evolution of policies and rules becomes critical, because if you think about it what should happen when changes apply to the policies? Does it imply resetting the learning process performed so far by the machines?!
- Europe has some complexity that may hinder AI adoption, but some of the complexity can also be a very valuable resource (for instance, multiple languages but integrated culture).
The take of Accenture
- companies are well aware that EU citizens are very careful about enterprise behaviour in the digital space
- there is still a massive issue related to digital literacy: data science is among the most requested skills, but plenty of people still need digital education.
Panel and speakers
As a side note: the panelists also addressed plenty of relevant questions from the audience, covering issues like training, granting access to the data also for small and medium enterprises and research institutions (what I call the data divide), breaking the monopoly on AI, and others. The panel was led by Andrea Renda (a research director at CEPS) and the people involved in the panel were:
- Lucilla Sioli, Artificial Intelligence and Digital Industry Director, DG Connect
- Eva Kaili, EU Parliament member
- Charina Chou, Global Policy Lead for Emerging Technologies, Google
- Andrew Sargison, Global Head of Advisory, Sales and Distribution Tech., Morgan-Stanley
- Rumman Chowdhury, Accenture
Wrapping up …
All of this defines a very clear context to the problem, but leaves plenty of space to create our own future.
It’s up to us as individuals, citizens, entrepreneurs, and companies to shape the future of this as a visionary, beneficial model for society, or as a dystopian future for humanity.