Brussels announces ‘GDPR’ for AI development

The European Commission is trying to tame the 'wild west' of AI development.
20 February 2020

European Commission’s President Ursula von der Leyen speaks during a press conference on AI. Source: AFP

Artificial Intelligence (AI) is gaining ground in industries rapidly, proving to be a disruptive force in the years to come. But is it spiraling out of control? 

According to the European Commission, progress is happening perhaps a little too rapidly and without restraint. 

Brussels this week outlined plans that could determine the fate of tech in Europe in new regulation. Industry Commissioner Thierry Breton suggested it would be comparable to the General Data Protection Regulation (GDPR).

A new AI regulatory paper outlines new governance for how tech companies operate in the region. The aim, it says, is to build a European society that is backed with digital solutions that “puts people first, opens up new opportunities for businesses, and boosts the development of trustworthy technology.”

Ursula von der Leyen, President of the European Commission, said: “ Today we are presenting our ambition to shape Europe’s digital future. It covers everything from cybersecurity to critical infrastructures, digital education to skills, democracy to media. I want that digital Europe reflects the best of Europe – open, fair, diverse, democratic, and confident.”

The gist of Europe’s digital layout

The Commission proposed an AI framework based on “excellence” and “trust.” By collaborating with the private and public sectors, the commission aims to mobilize resources along the entire value chain and implement the right procedures to deploy AI. It highlighted that the democratization of resources is vital to enable flexibility, scalability of businesses and to avoid the fragmentation of the single market.

 

In that regard, as well as seeking to protect personal data and privacy, strict rules will address unfair commercial practices. For high-risk cases involving sectors such as health, police, and transport, more profound transparency, traceability, and accountability of AI systems will be essential. 

This will be supervised by human inspectors; authorities will test and verify data used by algorithms. 

Even though the commission’s announcement marked a step forward in developing a stronger digital Europe, experts were disappointed with its lack of details in practice. It stated, for example, that businesses should approach AI development with an “Open, fair, diverse, democratic, and confident” attitude. 

However, this vagueness could be a cushion for tech companies to get in sync with the region’s new regulations gradually. 

Taking into account the ecosystem and unique role of AI in individual organizations, the proposed regulation aims to vocalize a common goal of AI adhering to regulations, without inhibiting its disruptive nature. 

To develop AI systems grounded in EU values would require the participation of all stakeholders, and that means clear guidelines from the off.

Cecilia Bonefeld-Dahl, Director General of DIGITALEUROPE, explained, “We need to continue the good cooperation between policymakers and industry and test different policies through pilots to ensure they achieve their goals.”

The Commission’s announcement may not sit well with tech giants as they will have to share their troves of data with smaller rivals. 

The EC remarked that dominant technology companies had a “data advantage” and were able to build empires by safeguarding their data.  EDiMA, representing platforms such as Google, Amazon, and Facebook, welcomed the proposals as a “starting point”,  stating it would engage with the European Commission to address some of the shared concerns.