EU in race to set global Artificial Intelligence ethics standards

"In the US, it is entirely driven by the private sector, large corporations, and some startups dealing with them. All the choices they will make are private choices that deal with collective values," warned Macron.

"On the other side, Chinese players collect a lot of data driven by a government whose principles and values are not ours," he pointed out.

Macron showed in the thought-provoking interview he knew that if you want to shape how AI will affect us, you have to be involved at the design stage, and set the rules.

"If we want to defend our way to deal with privacy, our collective preference for individual freedom versus technological progress, integrity of human beings and human DNA, if you want to manage your own choice of society, your choice of civilisation, you have to be able to be an acting part of this AI revolution," the French leader said, adding he wanted to "frame the discussion at a global scale".

A similar message came out of Brussels on Wednesday (25 April), when the European Commission presented its strategy paper on artificial intelligence - namely, that the EU should take the lead to shape the ethics of AI.

While there is no universal definition of AI, it generally refers to systems that can make (semi-)autonomous decisions based on analysing their environment or large data sets.

Autonomous cars are an obvious example, but many applications that have become mundane – like film suggestions by an on-demand platform or online translation services – were once seen as AI.

The commission paper – a 'communication' in EU jargon – listed three main goals: Boosting "the EU's technological and industrial capacity and AI uptake across the economy"; make sure there is "an appropriate ethical and legal framework" based on EU values; and "Prepare for socio-economic changes ".

It said that AI should respect the EU's "values and fundamental rights as well as ethical principles such as accountability and transparency".

"The EU can make a unique contribution to the worldwide debate on AI based on its values and fundamental rights," the document noted.

The paper said that the commission would set up a European AI Alliance that will be tasked to write draft AI ethics guidelines by the end of 2018.

This is faster than proposed in drafts. However, the commission originally had mulled a more formal name for the guidelines: a Charter on AI Ethics.

Such a charter or committee has also been suggested by other recent reports relating to AI.

"Our mission recommends the creation of a digital technology and AI ethics committee that is open to society," said a report on AI, written by a French mission led by mathematician and MP Cedric Villani.

The UK's House of Lords also recommended the development of an ethical code of conduct on AI.

"The public are entitled to be reassured that AI will be used in their interests, and will not be used to exploit or manipulate them, and many organisations and companies are as eager to confirm these hopes and assuage these concerns," wrote the UK's upper chamber of parliament.

They also recommended to Downing Street that it organised a global summit on AI in London by the end of 2019, and that the UK should take the lead on shaping the ethics of AI.

"The United States is unlikely to take this role," they wrote, because the Donald Trump administration appeared "relatively uninterested in AI", but also because the dominance of powerful US technology companies would make it less likely that "a truly democratic debate" would emerge there.

"Similarly, China shows few signs of wishing to limit the purview of the state or state-supported companies in utilising AI for alarmingly intrusive purposes," the British Lords wrote.

How to regulate AI

Before they wrote their report, the British politicians received oral and written evidence from more than 200 experts and organisations.

They wrote that while many witnesses argued that AI-specific laws should be adopted quickly, few of them actually knew what those laws should look like.

The Lords therefore said that AI-specific regulation would, for now, be "inappropriate".

The EU commission seems to have arrived at the same conclusion.

The draft versions of the strategy, as well as a source close to the commission, already showed confidence that already existing legislation could do the trick – although the commission noted it would monitor developments so that it could review EU legislation if needed.

One area where additional or adapted regulation may be needed, when it becomes legally blurred who is responsible for a decision with a negative outcome, is in the field of liability.

But when it comes to AI decisions based on personal data, new rules are already on their way.

In particular article 22 of the general data protection regulation (GDPR) that will kick in on 25 May prepares the ground, giving rights to 'data subjects' – us humans, when we are providing data.

"The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her," the regulation says – with exceptions listed in case the subject explicitly agrees to the automated processing or if it is required by law.

The lack of proposed legislation will likely disappoint MEPs.

The European Parliament had asked the commission in February 2017 to come with a proposal for a "legislative instrument on legal questions related to the development and use of robotics and AI foreseeable in the next 10 to 15 years".

This article was updated on 25 April 2018, 13:10, to replace citations from the strategy paper's drafts with citations from the final paper.