Saturday, May 05, 2018

Artificial Intelligence (Machine Learning) Ethics Guidelines for the European Union?

 China has advanced its much heralded approach to the management and direction of artificial intelligence (AI) since the Chinese State Council released the New Generation AI Development Plan (新一代人工智能发展规划的通知) in July 2017 (discussed here), grounded in economic and social development and support for national security.
Artificial intelligence has become a new focus of international competition. Artificial intelligence is a strategic technology that leads the future.  . . . Artificial intelligence has become a new engine for economic development. As the core driving force of the new round of industrial reforms, artificial intelligence will further release the tremendous energy accumulated from previous technological revolutions and industrial changes, and create new powerful engines to reconstruct economic activities such as production, distribution, exchange, and consumption.. . . Artificial intelligence brings new opportunities for social construction. China is now in the final phase of building a well-off society in an all-round way. Challenges such as population aging and resource and environmental constraints are still severe. Artificial intelligence is widely used in education, medical care, pensions, environmental protection, urban operations, and judicial services, which will greatly increase public awareness. (新一代人工智能发展规划的通知 machine translation)
The European Union has sought to put its own stamp on AI management at the governmental level. Its apoproach is, as might be expected, somewhat different.

"In its Communication "Artificial intelligence for Europe", the Commission puts forward a European approach to Artificial Intelligence based on three pillars:
  • Being ahead of technological developments and encouraging uptake by the public and private sectors
  • Prepare for socio-economic changes brought about by AI
  • Ensure an appropriate ethical and legal framework" (here)
The Communication and Fact Sheet follow. The Commission noted its desire to assemble a legitimating congress of stakeholders to focus on issues of AI around the approach it announced and consider relevant questions under the European AI Alliance. This Alliance is scheduled to be set up by July 2018, and there is a hope that an AI ethics guidelines will be published by the end of the year.


Useful links

The European Commission puts forward a European approach to artificial intelligence and robotics. It deals with technological, ethical, legal and socio-economic aspects to boost EU's research and industrial capacity and to put AI at the service of European citizens and economy.

Artificial intelligence (AI) has become an area of strategic importance and a key driver of economic development. It can bring solutions to many societal challenges from treating diseases to minimising the environmental impact of farming. However, socio-economic, legal and ethical impacts have to be carefully addressed.

It is essential to join forces in the European Union to stay at the forefront of this technological revolution, to ensure competitiveness and to shape the conditions for its development and use (ensuring respect of European values).
A European approach to Artificial Intelligence

In its communication, the European Commission puts forward a European approach to Artificial Intelligence based on three pillars:
Being ahead of technological developments and encouraging uptake by the public and private sectors

The Commission is increasing its annual investments in AI by 70% under the research and innovation programme Horizon 2020. It will reach EUR 1.5 billion for the period 2018-2020. It will:
connect and strengthen AI research centres across Europe;
support the development of an "AI-on-demand platform" that will provide access to relevant AI resources in the EU for all users;
support the development of AI applications in key sectors.

However, this represents only a small part of all the investments from the Member States and the private sector. This is the glue linking the individual efforts, to make together a solid investment, with an expected impact much greater than the sum of its parts.

Given the strategic importance of the topic and the support shown by the European countries signing the declaration of cooperation at the digital day, we can hope that Member States and the private sector will make similar efforts.

Joining forces at European level, the goal is to reach all together, more than EUR 20 billion per year over the next decade.

Prepare for socio-economic changes brought about by AI

To support the efforts of the Member States which are responsible for labour and education policies, the Commission will:
support business-education partnerships to attract and keep more AI talent in Europe;
set up dedicated training and retraining schemes for professionals;
foresee changes in the labour market and skills mismatch;
support digital skills and competences in science, technology, engineering, mathematics (STEM), entrepreneurship and creativity;
encourage Members States to modernise their education and training systems.
Ensure an appropriate ethical and legal framework

Some AI applications may raise new ethical and legal questions, related to liability or fairness of decision-making. The General Data Protection Regulation (GDPR) is a major step for building trust and the Commission wants to move a step forward on ensuring legal clarity in AI-based applications. In 2019 the Commission will develop and make available:
AI ethics guidelines;
Guidance on the interpretation of the Product Liability directive.

You can also consult the Staff Working Document for emerging digital technologies.
Declaration of cooperation on Artificial Intelligence

On 10 April 2018, 25 European countries signed a Declaration of cooperation on Artificial Intelligence. It builds further on the achievements and investments of the European research and business community in AI. The Commission will now work with Member States on a coordinated plan on AI to be delivered by the end of the year.
Background information

Artificial intelligence (AI) endows systems with the capability to analyse their environment and take decisions with some degree of autonomy to achieve goals.

Machine learning denotes the ability of a software/computer to learn from its environment or from a very large set of representative data, enabling systems to adapt their behaviour to changing circumstances or to perform tasks for which they have not been explicitly programmed.

To build robust models at the core of AI-based systems, high quality data is a key factor to improve performances. The Commission adopted a legislation to improve data sharing and open up more data for re-use. It includes public sector data as well as research and health data.
Useful links
Communication "Artificial Intelligence for Europe"
Declaration of cooperation on Artificial intelligence
Press release on Artificial intelligence: Commission outlines a European approach to boost investment and set ethical guidelines
Press release on creating European AI alliance
Staff Working Document: Liability for emerging digital technologies
Infographic on Artificial Intelligence
Questions and Answers on Artificial Intelligence, the way forward

Here is the E.U. A.I. Factsheet: "A European approach on AI will boost the European Union’s competitiveness and ensure trust based on European values. The European Commission has already invested significant amounts to bring benefits to our society and economy. This factsheet provides an overview of AI in Europe."


A European approach on Artificial Intelligence

Brussels, 25 April 2018
Questions and Answers
In today's Communication on Artificial Intelligence for Europe, the Commission is proposing a European approach to make the most out of the opportunities offered by artificial intelligence (AI), while addressing the new challenges AI brings. Building on European values, the Commission is proposing a three-pronged approach: increasing public and private investments; preparing for socio-economic changes brought about by AI; and ensuring an appropriate ethical and legal framework.

I - Boosting the EU's technological and industrial capacity and AI uptake across the economy
What kind of challenges can AI address? What kind of AI projects will the EU fund?
AI helps us solve many societal challenges from helping doctors make faster and more accurate medical diagnoses to assisting farmers in using fewer pesticides for their crops. It also helps public administrations to provide tailor-made responses to citizens and to decrease the number of traffic accidents. AI can help fight climate change or anticipate cybersecurity threats. The Commission will fund projects to support the use of AI in many applications, from health to transport, and to digitise industry. EU funding will also support projects to improve the performance of AI technology (e.g. the quality of speech recognition).
The Commission will support fundamental research, and also help bring more innovations to the market through the European Innovation Council pilot. Additionally, the Commission will support Member States' efforts to jointly establish AI research excellence centres across Europe. The goal is to encourage networking and collaboration between the centres, including the exchange of researchers and joint research projects.
The Commission will also support the uptake of AI across Europe, with a toolbox for potential users, focusing on small and medium-sized enterprises, non-tech companies and public administrations. The set of measures will include an EU 'AI-on-demand platform' giving advice and easy access to the latest algorithms and expertise; a network of AI-focused Digital Innovation Hubs facilitating testing and experimentation; and industrial data platforms offering high quality datasets. Several priorities have also been identified for the post-2020 multiannual financial framework (such as increased support in fields such as explainable AI to develop AI systems in a way which allows humans to understand the basis of their action or AI systems which need less data).
How will the European Fund for Strategic Investments (EFSI) help companies to adopt AI and when?
The European Fund for Strategic Investments will support the development and the uptake of AI, as part of the wider efforts to promote digitisation. The Commission – together with its strategic partner, the European Investment Bank Group – aims to mobilise more than €500 million in total investments in the period 2018-2020 across a range of key sectors. To this end, a thematic investment platform under the EFSI could be set up. In addition, the European Commission and the European Investment Fund have just launched VentureEU, a €2.1 billion Pan-European Venture Capital Fund-of-Funds programme, to boost investment in innovative start-up and scale-up companies across Europe.
What are Digital Innovation Hubs and how will they contribute to the use of AI?
Digital Innovation Hubs are local ecosystems that help companies in their vicinity (especially small and medium-sized enterprises) to take advantage of digital opportunities. They offer expertise on technologies, testing, skills, business models, finance, market intelligence and networking. For example, a small company that produces metal parts for the automotive industry could consult the regional hub and ask for advice on how to improve the manufacturing process with AI. Experts from the hub would then visit the factory, analyse the production process, consult with other AI experts in the network of hubs, make a proposal and then implement it. These activities would be partially financed with EU money.

II - Preparing for socioeconomic changes
What is the Blueprint for Sectoral cooperation on Skills? Which sectors are targeted?
Europeans should have every opportunity to acquire the skills and knowledge they need and to master new technology. National schemes are essential for providing such up-skilling and training. They can benefit from support by the European Structural and Investment Funds (supporting skills development with €27 billion over the period 2014-2020, out of which the European Social Fund invests €2.3 billion specifically in digital skills) and should also benefit from support from the private sector.
The Blueprint for Sectoral cooperation on Skills identifies skills needed and gaps in a sector and connects with partners that can help address those needs by developing a common European strategy and curricula development. Five sectors were chosen to pilot the Blueprint in 2017 (automotive, maritime technology, space/geo information, textile/leather/clothing/footwear and tourism) and six other sectors have been recently added (additive manufacturing, construction, maritime shipping, paper-based value chain, renewable energy and green technologies and steel industry) with EU funding support of close to €50 million.
What is the Digital Opportunity Traineeship in advanced digital skills for students and recent graduates? How will it support AI?
The Digital Opportunity traineeship initiative will provide cross-border traineeships for up to 6,000 students and recent graduates as of summer 2018. It will give students of all disciplines the opportunity to get hands-on digital experience within companies, in fields demanded where there is a skills gap, and strengthen their ICT skills, in fields such as AI.
In addition to the Digital Opportunity traineeships, the Commission asked all Member States to develop national digital skills strategies by mid-2017 and to set up national coalitions to support their implementation. National Coalitions bring together ICT and ICT-intensive companies, education and training providers, education and employment ministries, public and private employment services, associations, non-profit organisations and social partners, who all develop measures to bring digital skills to all levels of society. Through the Digital Skills and Jobs Coalition the Commission will encourage business-education partnerships for AI.
The European Institute of Innovation & Technology also designs specific programmes at Master and PhD levels to address needs arising from the digital sector and digital transformation. The programmes combine in-depth technical skills with strong innovation and entrepreneurial components. They develop skills linked to data collection techniques, data analysis methods, computer science, electronic engineering, deep learning and image recognition. These are all skills needed in areas of AI applications such as self-driving cars and robotics and image/video identification with applications in security and safety.

III - Ensuring an appropriate ethical and legal framework
How is the Commission encouraging the transparency of algorithms?
Algorithms are behind more and more decisions that affect our everyday lives such as access to universities, getting a loan, or the selection of filtering of information; transparency is therefore crucial. In several areas, there are already EU rules for algorithmic decisions. Examples include automated decisions based on personal data (General Data Protection Regulation, GDPR) and for high-frequency trading on the stock-market (Markets in Financial Instruments Directive, MiFID II).
Algorithmic transparency will be a topic addressed in the AI ethics guidelines to be developed by the end of the year. The AI ethics guidelines will build on work from various relevant initiatives such as the Algorithmic Awareness Building Project which will address issues related to algorithmic transparency, accountability and fairness.
Algorithmic transparency is not about disclosure of source code as such. It can take different forms, depending on the situation, including meaningful explanation (as required in GDPR), or reporting to the competent authorities (as required in MiFID II).
What is the product liability directive? Why is guidance needed?
The EU has liability rules for defective products. The Product Liability Directive dates from 1985 and strikes a careful balance between protecting consumers and encouraging businesses to market innovative products. The Directive covers a broad range of products and possible scenarios.
In principle, if AI is integrated into a product and a defect can be proven in a product that caused material damage to a person, the producer will be liable to pay compensation.
The actual cause of events that lead to damage or incident is decisive for the attribution of liability. The Commission plans to issue an interpretative guidance clarifying concepts of the Directive in view of the new technologies, building on a first assessment on liability for emerging digital technologies published today.
How does the General Data Protection Regulation apply to AI?
The General Data Protection Regulation (GDPR) ensures a high standard of personal data protection, including the principles of data protection by design and by default. It has provisions on decision-making based solely on automated processing, including profiling (AI-based systems). In such cases, data subjects have the right to be provided with meaningful information about the logic involved in the decision.
The GDPR also gives individuals the right not to be subject solely to automated decision-making (except in certain situations) such as automatic refusal of an online credit application or e-recruiting practices without any human intervention. Such processing includes profiling that consists of any form of automated processing of personal data evaluating the personal aspects relating to a natural person (AI-based systems), in particular to analyse or predict aspects concerning the data subject's performance at work, economic situation, health, personal preferences or interests, reliability or behaviour, location or movements, where it produces legal effects concerning him or her or similarly significantly affects him or her.
What will the ethics guidelines be about? What role will the AI Alliance play?
Draft AI ethics guidelines will be developed on the basis of the EU's Charter of Fundamental Rights, following a large consultation of stakeholders within the AI Alliance. The draft guidelines will build on the statement published by the European Group of Ethics in Science and New Technologies. They will address issues such as the future of work, fairness, safety, social inclusion, algorithmic transparency, and more broadly, will examine the impact on fundamental rights, including privacy, dignity, consumer protection and non-discrimination.
Given the scale of the challenge associated with AI, the full participation of all actors including businesses, academics, consumer organisations, trade unions, policy makers and representatives of civil society is essential. This is why the Commission wants to bring together a broad community of stakeholders around AI-relevant questions under the European AI Alliance. The Alliance will be set up by July 2018, and AI ethics guidelines will be published by the end of the year.
For More Information
Press release

No comments: