EU could make AI knowledge transfer obligatory

Data-sharing could help keep up Europe’s pace in the global AI race.
27 June 2019

European headquarters in Brussels. Source: Shutterstock

Artificial Intelligence (AI) is a hotly-contested market worldwide. Companies that can develop and deploy their technology at scale hold the key to their own success. 

With Europe at risk of being outpaced by US-China arms race, however, a panel of EU Commission-backed experts recommend introducing AI knowledge-transfer regulations in order to maintain a competitive market and to boost the technology’s growth. 

With the recommendations likely to influence future policies, companies based in the EU market could be obliged to share baseplate AI technology with rival firms. 

The line came as one of a batch of policy recommendations for ‘Trustworthy AI’ by the High-Level Expert Group on Artificial Intelligence (AI HLEG), a panel of 52 academic and business figures in AI, technology, engineering, and law. Members include Google, IBM, Orange, SAP, AXA, Bayer, and Santander.  

AI HLEG said it was important European industries adopt AI in order to adapt to the demands of a “rapidly evolving digital economy.” 

It noted that in 2018, “three-quarters of European businesses did not adopt AI in any shape or form.” Meanwhile, just 2-3 percent of companies incorporated AI across their whole organization. 

The report urged policymakers to foster transformation, particularly among SMEs which account for 56 percent of Europe’s total turnover. “[…] it is likewise important that larger companies in Europe accelerate their transformation,” the report states. 

“A multi-stakeholder approach, bringing all actors around the table to join forces, is crucial to foster sectoral AI ecosystems where most of the transformative power of AI and sustainable growth is expected.” 

Should AI vendors share knowledge?

A key driver of the wide adoption of AI technology, according to AI HLEG, will be in technology transfer: “Consider the introduction of a data access regime on FRAND terms, namely fair, reasonable, and non-discriminatory,” the group stated.  

“In addition, data interoperability amongst market players should be incentivized and required, where appropriate.”

As noted by the publication Engineering & Technology, an example of how a lack of data sharing in AI can impact sector competition can be seen in the virtual assistants available for smart homes and devices by leading tech firms. 

Owed to Google and Microsoft’s access to search data, the Google Assistant and Microsoft Cortana assistants were able to give better feedback than Apple’s Siri (a now privacy-focused company) and Amazon’s Alexa (a retail-based firm). 

Ownership of data in this respect can lead certain companies to become dominant in their sector— other firms and new entrants wouldn’t always be able to compete. 

Beyond AI knowledge transfer, other recommendations included freeing up more budget for funding for the development of ethical AI— including Innovation Hubs, loans, legal support, education, and training.