Patent pools, AI and the Digital Single Market 

The digital world is set to be transformed by the rapid advances we are now seeing in AI technology. A new book spotlights the role patent pools can play in ensuring Europe makes the most of this 

 By Jacob Schindler 

For nearly a decade, the European Union has pursued the Digital Single Market strategy. This has been designed to improve access to digital goods and services, create a favourable environment for digital networks and harness digital transformation as an engine of European growth.  

With rapid advancements in AI poised to re-shape the digital world, a group of legal and policy researchers has published a new academic volume, Digital Single Market and Artificial Intelligence: AI Act and Intellectual Property in the Digital Transition. Given the EU AI Act became law last month, the book offers a timely look at a wide range of intersections between AI and EU policy. 

“The virtuous circle of licenses in standard essential patents”, a chapter in the new collection authored by Sisvel Video Coding Platform programme manager Valentina Piola, explores the potential role for patent pools in this new era of AI-defined innovation. In another chapter, “AI and standard essential patents: the MPAI experience", Sisvel founder Roberto Dini and the president of MPAI, Leonardo Chiariglione, examine the standards development organisation’s unique approach to facilitating SEP licensing. 

Licensing levels the playing field 

As is evident from a range of other technologies underlying the digital transition, notably 5G, Wi-Fi and IoT, global standards play a fundamental role in enabling the interoperability that promotes the broad access and level playing field envisioned by the Digital Single Market.  

The licensing of standard essential patents on FRAND terms, in turn, rewards innovators who contribute to the creation of these building blocks of digital transformation. This gives rise to the Inventive Loop, which sees a sustainable proportion of profits and royalties ploughed right back into research and development. 

As Piola points out, this virtuous cycle is especially important for encouraging SMEs, universities, research institutes and startups to play a role in shaping foundational technologies and standards. “It goes without saying that these innovators cannot count on the same level of expenditure as the big tech companies. Therefore, licensing is the key to creating a sustainable ecosystem for innovation,” she writes. 

AI-related standards emerging 

In the AI domain, R&D investments and patenting activity are both picking up speed. When it comes to the flagship AI products that have gained so much attention over the last year, most Big Tech companies are pursuing walled garden strategies. But in some areas of AI, open standardisation initiatives are emerging.  

One of the most recently created standardisation bodies, and the focus of the chapter by Dini and Chiariglione, is the Moving Picture, Audio and Data Coding by Artificial Intelligence (MPAI) community, founded in 2020. This describes itself as an “international, unaffiliated, non-profit organisation with the mission to develop Artificial Intelligence (AI) enabled data coding specifications, with clear Intellectual Property Rights (IPR) licensing frameworks”. Among other activities, it has developed nine technical specifications, in areas which include AI app execution, enhanced audio and connected autonomous vehicles. 

MPAI has adopted a novel IP management model known as the FrameWork Licence or FWL. Developed in response to challenges under the existing paradigm, including holdout and the weakening of the one-stop-shop offering in some technology fields, MPAI’s model aims to help SEP holders and implementers reach agreements faster and with less friction. As the name suggests, the FrameWork Licence sets out parameters for negotiations rather than exact commercial terms.  

The MPAI effort shows that stakeholders in the standards world are thinking carefully about how to create licensing models that are fit for purpose in the AI era. 

Pool planning should start early 

It is critical to consider patent pools from the beginnings of such an initiative in a pre-commercial process that precedes formal pool facilitation undertaken by an administrator, Piola says. Ideally, licensing frameworks should also be agreed at an early stage to provide clarity and reduce market uncertainty.  

As Dini and Chiariglione explain, MPAI differs from some other standards initiatives in that its statutes explicitly call for early adoption of a framework licence for each of its standards. This proactive step is intended to ensure that implementers have a clear picture of licensing terms and aggregate royalty rates from the outset. 

Other keys to a successful pool highlighted by Piola include:  

• balance between patent owners and implementers; 

• appeal to SME, university, and startup rightsholders; 

• independence of the administrator; and  

• rigorous third-party essentiality checking provisions. 

In sum, there is a clear role for independent, transparent and balanced patent pools to play in this ascendent area of technology. As in so many other fields, it is right in line with European policymakers’ goals for shaping the bloc’s digital future.  

As Piola points out in her contribution to this important publication - and Sisvel’s head of licensing programmes David Muus has highlighted elsewhere - the European Commission has recognised this consistently in the past. It’s something they should revisit very carefully as they plot the next steps for the Digital Single Market in the AI era.   

Jacob Schindler is senior manager in Sisvel’s Content and Strategic Communications team. He is based in Hong Kong