The EU sets out a “European” vision for AI

The European Commission has kick-started its campaign to lead the world when it comes to standards and ethical guidelines for artificial intelligence.

This is the precarious position the AI world currently finds itself in; setting the standards which will govern development for years to come. In the US, the reigns have been firmly grasped by the private sector, while over in Asia, China is leading the charge, heavily influenced by government ambitions. These two sets of ambitions could not be ideologically further apart, though perhaps the European Commission might be able to create an effective middle-ground position.

“Just as the steam engine and electricity did in the past, AI is transforming our world,” said Vice-President for the Digital Single Market Andrus Ansip.

“It presents new challenges that Europe should meet together in order for AI to succeed and work for everyone. We need to invest at least €20 billion by the end of 2020. The Commission is playing its part: today, we are giving a boost to researchers so that they can develop the next generation of AI technologies and applications, and to companies, so that they can embrace and incorporate them.”

AI is still an embryonic technology, or umbrella of technologies. While machine learning has been hitting the headlines for months now, this is only one aspect of the complicated web of intelligence. Computer perception, computational creativity, decision making capabilities and genuine personalisation are still in the pipeline, making now a good time for the European Commission to weigh in. In a strange turn of events, the boresome bureaucrats showing ambitions to get involved in the development of a technology at the right time.

While it might only be a PR-quip, Ansip is right; AI will redefine the world we live in. Perhaps this is why it is important the even-handed (in comparison) and privacy conscious European regulator needs to aggressively assert itself as a key arbitrator in the technology’s development. An even balance needs to be struck between commercial objectives of the private sector, the often nefarious actions of government and the well-being of the consumer. It will be a delicate compromise, but for European principles of privacy and data protection to be upheld, the European Commission cannot afford to mess this up.

The ‘European’ approach to the AI conundrum will see a three-pronged approach which builds on ‘European’ values. The three areas are as follows; increasing public and private investments, preparing for socio-economic changes brought about by AI and ensuring an appropriate ethical and legal framework.

Looking at the investment side of things, the European Commission has said it will commit roughly €2.6 billion over the duration of Horizon 2020 campaign to AI-related projects. This figure is expected to trigger an additional €2.5 billion of funding from existing public-private partnerships, while the European Fund for Strategic Investments will be mobilised to provide companies and start-ups with additional support to invest in AI. The idea here is to create an environment which aids the development of the technology.

One aspect of creating this environment will be to increase the availability of data in the EU. This includes data from public utilities and the environment as well as research and health data. Alongside the three-pronged strategy, the Commission is also introducing new proposals to ease access to and reusability of public sector data, increase scientific data sharing and create a more collaborative environment. This is perhaps an important step in removing the grey areas on the legal and technical principles that should govern data sharing collaboration in the public and private sectors.

Secondly, preparing for socio-economic changes brought about by AI. When you look at the development of AI, some jobs will be made, more will be lost and a huge number will be transformed to something new. This area is a simple one to define, but incredibly difficult to action. The European Commission will have to prod the lethargic legislators in the member states into action to modernise their education and training systems and support labour market transitions.

Seeing as very few governments have actually recognised the pain which will be experienced through the implementation of AI, this will certainly be a difficult one. As it stands, most politicians are happy to discuss the opportunities and rewards, but few want to discuss how many people will be made redundant. Not enough is being done to prepare the workforce and considering the rate of change and adoption, it is a very worrying trend. First and foremost, governments have to recognise that AI will not be everything good to everyone, then work can begin in supporting the transition of these individuals and businesses.

The final area will be even more difficult. New technologies pose new questions and challenges to regulations, therefore a new framework will need to be developed to address the new legal and ethical dilemmas posed by AI. The Commission will present ethical guidelines on AI development by the end of 2018, based on the EU’s Charter of Fundamental Rights, taking into account principles such as data protection and transparency.

“The development of the artificial intelligence industry will be stifled without the right legal framework. Algorithm accountability for when things go wrong is one of the key issues,” said Mark Deem, Partner at law firm Cooley. “The problem is that liability in artificial intelligence, especially machine learning, is far more complex compared to most other environments.

“Given the multiple participants involved in any artificial intelligence value chain, an efficient way of solving this liability problem will be for market players to define the parameters of their own liabilities. This will provide assurance throughout the industry and enable the development of insurance products to carry liability, which will act as a catalyst to investment. The approach of the European Commission therefore to seek to bring industry, government and academic insight together is the right one.”

Europe is generally seen as lagging behind the US and Asia when it comes to technology development and implementation, though considering these are still the early days of AI, there is still time for Europe to stake a claim. For this to be a possibility, the European Commission will have to be aggressive and efficient when it comes to developing AI frameworks and developing the right environment. Form is not on the side of the cumbersome bureaucrats, but it is an important mission for Brussels.