A European Approach to AI 2020
A Summary of The EU White Paper on Artificial Intelligence February 2020
When I heard that the EU Commission was releasing a white paper on artificial intelligence I had to read it. So this is a short collection of excerpts from the recent report called On Artificial Intelligence — A European approach to excellence and trust.
Most of the Bold text in this article is emphasised by me and does not occur in the original document.
The report starts with a clear twin objective of the EU.
The Twin Objective
The purpose of this White Paper is to set out policy options on how to achieve these two objectives:
- Promoting the uptake of AI.
- Addressing the risks associated with certain uses of this new technology.
Why are they doing this?
“Against a background of fierce global competition, a solid European approach is needed, building on the European strategy for AI presented in April 20181 . To address the opportunities and challenges of AI, the EU must act as one and define its own way, based on European values, to promote the development and deployment of AI. The Commission is committed to enabling scientific breakthrough, to preserving the EU’s technological leadership and to ensuring that new technologies are at the service of all Europeans — improving their lives while respecting their rights.”
Sustainable Growth and Societal Wellbeing
The outline and the foremost imperative is twofold, and there is a mentioned importance on managing data.
Europe’s current and future sustainable economic growth and societal wellbeing increasingly draws on value created by data. AI is one of the most important applications of the data economy. Today most data are related to consumers and are stored and processed on central cloud-based infrastructure.
What is AI According to the Report?
Most reports have different perspectives on what AI is and the descriptive terms used to define the field.
Simply put, AI is a collection of technologies that combine data, algorithms and computing power.
The strategy mentions three groups:
- Citizens (healthcare, services and products).
- Business development (machinery, transport, cybersecurity, farming, the green and circular economy, healthcare and high-value added sectors like fashion and tourism).
- Public interest (reducing the costs of providing services transport, education, energy and waste management) – improving the sustainability of products.
The Main Building Blocks of the White Paper
Early in the report there is a mention of a few building blocks.
This White Paper presents policy options to enable a trustworthy and secure development of AI in Europe, in full respect of the values and rights of EU citizens
- The policy framework setting out measures to align efforts at European, national and regional level. In partnership between the private and the public sector, the aim of the framework is to mobilise resources to achieve an ‘ecosystem of excellence’ along the entire value chain, starting in research and innovation, and to create the right incentives to accelerate the adoption of solutions based on AI, including by small and medium-sized enterprises (SMEs).
- The key elements of a future regulatory framework for AI in Europe that will create a unique ‘ecosystem of trust’. To do so, it must ensure compliance with EU rules, including the rules protecting fundamental rights and consumers’ rights, in particular for AI systems operated in the EU that pose a high risk . Building an ecosystem of trust is a policy objective in itself, and should give citizens the confidence to take up AI applications and give companies and public organisations the legal certainty to innovate using AI. The Commission strongly supports a human-centric approach based on the Communication on Building Trust in Human-Centric AI and will also take into account the input obtained during the piloting phase of the Ethics Guidelines prepared by the High-Level Expert Group on AI.
There is a mention too of securing data to ensure ‘Europe’s technological sovereignty’.
The European strategy for data, which accompanies this White Paper, aims to enable Europe to become the most attractive, secure and dynamic data-agile economy in the world — empowering Europe with data to improve decisions and better the lives of all its citizens. […] Harnessing the capacity of the EU to invest in next generation technologies and infrastructures, as well as in digital competences like data literacy, will increase Europe’s technological sovereignty in key enabling technologies and infrastructures for the data economy.
AI Value Chain
An important mention in my opinion is the mention of the value chain relating to artificial intelligence in the communication.
Europe should leverage its strengths to expand its position in the ecosystems and along the value chain, from certain hardware manufacturing sectors to software all the way to services. This is already happening to an extent.
Investment in AI Research
Investment in AI has increased drastically in the EU, yet in terms of the size it is lagging behind other regions.
Over the past three years, EU funding for research and innovation for AI has risen to €1.5 billion, i.e. a 70% increase compared to the previous period. […] However, investment in research and innovation in Europe is still a fraction of the public and private investment in other regions of the world. Some €3.2 billion were invested in AI in Europe in 2016, compared to around €12.1 billion in North America and €6.5 billion in Asia.
Data Produced and Data Centres
As there is a drastic increase in data produced EU needs to adapt its strategy accordingly.
The volume of data produced in the world is growing rapidly, from 33 zettabytes in 2018 to an expected 175 zettabytes in 2025. Each new wave of data brings opportunities for Europe to position itself in the data-agile economy and to become a world leader in this area. Furthermore, the way in which data are stored and processed will change dramatically over the coming five years. Today 80% of data processing and analysis that takes place in the cloud occurs in data centres and centralised computing facilities, and 20% in smart connected objects, such as cars, home appliances or manufacturing robots, and in computing facilities close to the user (“edge computing”). By 2025 these proportions are set to change markedly .
Low-Power Electronics and Neuromorphic Solutions
There is a mention of a word that to some extent seems unknown, and a valid point made that all applications need to be energy efficient.
Europe is a global leader in low-power electronics which is key for the next generation of specialised processors for AI. This market is currently dominated by non-EU players. This could change with the help of initiatives such as the European Processor Initiative, which focuses on developing low-power computing systems for both edge and next generation high-performance computing, and the work of the Key Digital Technology Joint Undertaking, proposed to start in 2021. Europe also leads in neuromorphic solutions that are ideally suited to automating industrial processes (industry 4.0) and transport modes. They can improve energy efficiency by several orders of magnitude.
There is talk in policy circles relating to technology of quantum supremacy, and it seems the EU holds a clear interest in this area as well.
Recent advances in quantum computing will generate exponential increases in processing capacity. Europe can be at the forefront of this technology thanks to its academic strengths in quantum computing, as well as European industry’s strong position in quantum simulators and programming environments for quantum computing. European initiatives that aim to increase the availability of quantum testing and experimentation facilities will help apply these new quantum solutions to a number of industrial and academic sectors
Ecosystem of Excellence
I have listed the different action points and abbreviated heavily. I apologise for any omissions or misunderstandings this may cause.
Action 1: working with member states on AI. Revision of the Coordinated Plan to be adopted by end 2020
The objective is to attract over €20 billion of total investment in the EU per year in AI over the next decade.
To stimulate private and public investment, the EU will make available resources from:
- The Digital Europe Programme,
- Horizon Europe
- European Structural and Investment Funds
Action 2: Focused efforts on the research and innovation community.
- More synergies in the networks
- Excellence and testing centres
- The updated Digital Education Action Plan will help make better use of data and AI-based technologies such as learning and predictive analytics with the aim to improve education and training systems and make them fit for the digital age.
- A lighthouse focused on AI to coordinate efforts “…a lighthouse centre of research and innovation for AI in Europe would attract talent from all over the world due to the possibilities it could offer.”
Action 3: offer world-leading masters programmes in AI (Digital Europe Programme) and attract the best professors.
It will also be important to ensure that SMEs can access and use AI. To this end, the Digital Innovation Hubs and the AI-on-demand platform should be strengthened further and foster collaboration between SMEs. […] the Commission plans to further scale up access to finance in AI under InvestEU . AI is explicitly mentioned among the eligible areas for the use of the InvestEU guarantee.
Action 4: the Commission will work with Member States to ensure that at least one digital innovation hub per Member State has a high degree of specialisation on AI
Action 5: the Commission will set up a new public private partnership in AI
Action 6: ‘Adopt AI programme’ that will support public procurement of AI systems, and help to transform public procurement processes themselves.
Securing Access to Data and Computing Infrastructures
Furthermore there is funding to support expansion of digital infrastructure in Europe (although this is dwarfed to some extent by big tech private spending), however support by EU financed further by private initiatives in the EU could be a possible fascinating combination.
The Commission has proposed more than €4 billion under the Digital Europe Programme to support high-performance and quantum computing, including edge computing and AI, data and cloud infrastructure. The European data strategy develops these priorities further.
There is a brief mention of the 2030 agenda towards the end of section 4.H:
It is also clear that the responsible development and use of AI can be a driving force to achieve the Sustainable Development Goals and advance the 2030 Agenda
Guidelines on Trustworthy AI
The whitepaper mentions the guidelines created in 2019 and talks about the follow-up.
During the second half of 2019, over 350 organisations have tested this assessment list and sent feedback. The High-Level Group is in the process of revising its guidelines in light of this feedback and will finalise this work by June 2020. A key result of the feedback process is that while a number of the requirements are already reflected in existing legal or regulatory regimes, those regarding transparency, traceability and human oversight are not specifically covered under current legislation in many economic sectors.
They mention two common faultlines (my word) in AI:
- Certain AI algorithms, when exploited for predicting criminal recidivism, can display gender and racial bias, demonstrating different recidivism prediction probability for women vs men or for nationals vs foreigners. Source: Tolan S., Miron M., Gomez E. and Castillo C. “Why Machine Learning May Lead to Unfairness: Evidence from Risk Assessment for Juvenile Justice in Catalonia”, Best Paper Award, International Conference on AI and Law, 2019
- Certain AI programmes for facial analysis display gender and racial bias, demonstrating low errors for determining the gender of lighter-skinned men but high errors in determining gender for darker-skinned women. Source: Joy Buolamwini, Timnit Gebru; Proceedings of the 1st Conference on Fairness, Accountability and Transparency, PMLR 81:77–91, 2018.
They mention a few specific risks the commission wishes to address, I have added the bold text:
- Effective application and enforcement of existing EU and national legislation: the key characteristics of AI create challenges for ensuring the proper application and enforcement of EU and national legislation.
- Limitations of scope of existing EU legislation: an essential focus of EU product safety legislation is on the placing of products on the market.
- Changing functionality of AI systems: the integration of software, including AI, into products can modify the functioning of such products and systems during their lifecycle.
- Uncertainty as regards the allocation of responsibilities between different economic operators in the supply chain.
- Changes to the concept of safety: the use of AI in products and services can give rise to risks that EU legislation currently does not explicitly address.
It mentions too the report on the safety and liability implications of Artificial Intelligence, the Internet of Things and robotics.
A Future Regulatory Framework
A key issue for the future specific regulatory framework on AI intelligence is to determine the scope of its application. […] In any new legal instrument, the definition of AI will need to be sufficiently flexible to accommodate technical progress while being precise enough to provide the necessary legal certainty.
- First, the AI application is employed in a sector where, given the characteristics of the activities typically undertaken, significant risks can be expected to occur. (healthcare; transport; energy and parts of the public sector)
- Second, the AI application in the sector in question is, in addition, used in such a manner that significant risks are likely to arise.
Types of legal framework, as requirements in these areas:
- Training data.
- Keeping of records and data.
- Information provision
- Robustness and accuracy
- Human oversight
- Specific requirements for remote biometric identification
Still there is a mention of: voluntary labelling for no-high risk AI application.
This is an attempt of shortening or taking a few excerpts from the whitepaper.
I hope it was useful and I recommend you as always to read the paper to get a better overview in full if you have the time.
This is #500daysofAI and you are reading article 264. I am writing one new article about or related to artificial intelligence every day for 500 days. My current focus for 100 days 200–300 is national and international strategies for artificial intelligence.