Share
Maximising the benefits of artificial intelligence through future-proof rules on Text and Data Mining
Dear President Juncker,
Dear Vice-President Ansip, Dear Vice-President Katainen,
Dear Commissioner Bieńkowska, Dear Commissioner Gabriel,
We, members of 23 organisations representing universities, large and small technology companies, telecommunications and Internet service providers, startups and scaleups, libraries, scientific and research funding and performing organisations, open access publishers, investigative and data journalists and non-profits, fully support the shared ambitions of the European Commission and the Council of the EU to stimulate the development and boost investment into Artificial Intelligence (AI) in Europe.
As rightly acknowledged by the European Commission[1], AI, machine learning and data analytics offer vast potential for economic growth and increased competitiveness in Europe[2] and their benefits in healthcare diagnostics, supply chain management and sustainable energy management cannot be underestimated. Yet, the uptake of AI, machine learning and data analytics in Europe remains limited. In 2017, only 10% of SMEs in Europe used big data analytics[3]. European companies currently lag behind their US and Chinese competitors in this area. The time is ripe for the European Commission to come up with the right regulatory and policy framework to encourage European companies to become leaders in AI.
We are deeply concerned by the fact that the draft Communication on AI does not recognise the foundational role that Text and Data Mining (TDM) plays in AI. TDM is a building block for both machine and deep learning: without the ability of computers to analyse very large amounts of data, employ cognitive technologies to allow the learning of patterns, AI is not possible.
As such, we fail to comprehend why, although the Commission accepts the fact that AI and machine learning need vast amounts of data for training, it is currently pushing for a restrictive TDM exception (article 3) in the Copyright Directive that will remove the ability for startups, businesses, or public-private collaborators to use TDM to develop any of the innovations it seeks to foster, because the proposed TDM exception only covers non-for-profit and public research institutions. Limiting the ability of private companies to carry out TDM in Europe will inevitably lead to the most promising European AI startups and companies relocating to the US[4] or Japan[5], where they will have access to broader datasets, and will be able to build algorithms of better quality. As it stands, article 3 of the Copyright Directive will have the opposite effect of what the Commission wishes to achieve with its AI Strategy.
We would also like to point out that the ongoing efforts of the European Commission on the European Open Science Cloud are playing an important role in safeguarding sustainable access to datasets and thus are a key enabling factor for progress in AI. These efforts and considerable investments should not be endangered.
We are convinced that any comprehensive strategy to make Europe globally competitive in the race to develop and implement AI-powered solutions must include robust support for TDM in both the public and private industries. The quality of AI algorithms depends on access to large and diverse data-sets. This is recognised for instance in the USA and Japan, where TDM activities are broadly allowed and even encouraged.
As outlined in the recommendations of the recent AI report of French MP Cédric Villani[6], there is an urgent need to promote TDM practices to make Europe a global leader in AI. As the European Commission and the Bulgarian Presidency of the Council of the EU are preparing discussions around AI during the upcoming Digital Day in Brussels on 10 April and ahead of the publication of the Commission’s AI strategy on 25 April, we call on all European institutions and in particular the European Commission to take a broader view on Artificial Intelligence and Text and Data Mining, and adopt a broader TDM exception in the copyright directive that will allow for a high-quality, thriving AI industry in Europe.
[2] Why Artificial Intelligence is the Future of Growth, Mark Purdy and Paul Daugherty, Accenture, 2017
[3] 1 in 10 EU businesses analyses big data, http://ec.europa.eu/eurostat/web/products-eurostat-news/-/EDN-20170516-1?inheritRedirect=true&redirect=/eurostat/
[4] The Exception for Text and Data Mining (TDM) in the Proposed Directive on Copyright in the Digital Single Market – Technical Aspects, Eleonora Rosati
[5] The Exception for Text and Data Mining (TDM) in the Proposed Directive on Copyright in the Digital Single Market – Legal Aspects, Christophe Geiger, Giancarlo Frosio and Oleksandr Bulayenko
[6] https://www.aiforhumanity.fr/pdfs/MissionVillani_Report_ENG-VF.pdf
—
Download the letter
OpenLetter to European Commission on AI and TDM_9April2018
—
Signatories
- Aalto University;
- Allied for Startups;
- BSA | The Software Alliance;
- Bitkom;
- Center for Democracy and Technology (CDT);
- Computer and Communications Industry Association (CCIA);
- Confederation of Open Access Repositories (COAR);
- Copyright 4 Creativity (C4C);
- DIGITALEUROPE;
- EDiMA;
- ESOMAR;
- European Alliance for Research Excellence (EARE);
- European DIGITAL SME Alliance;
- European University Association (EUA);
- European Bureau of Library, Information and Documentation Associations (EBLIDA);
- Frontiers;
- League of European Research Universities (LERU);
- Ligue des Bibliothèques Européennes de Recherche – Association of European Research Libraries (LIBER);
- Open State Foundation;
- Research Libraries UK (RLUK);
- Society of College, National and University Libraries (SCONUL);
- SPARC Europe;
- UCL (University College London).