Artificial Intelligence (AI) will fundamentally impact how we work, communicate, travel and live in the future. The technology is already an integral part of our daily lives : playlists on Spotify, Amazon’s product recommendations, fraud prevention or Google Maps – most of the digital facilitators are powered by AI. But how to ensure we learn about our machines while these machines are learning for us?
According to the Oxford Internet Institute, people hesitate to put their personal lives in the hands of an AI assistant, especially when the decision-making process is not sufficiently transparent. As the algorithmic decision-making process is complex and differs from sector to sector, information is generally less accessible to a wider audience. However, the impact of AI will be of everybody’s concern. At the same time, digital education is unevenly distributed across Europe and stark knowledge gaps persist. How, then, can citizens become more knowledgeable about the benefits and risks arising from AI technology?
Transparency is a key principle emphasised in the European Commission’s Guidelines for Trustworthy AI. As such, the principle foregrounds transparent technology and use of AI:
- explainability of the algorithmic decision-making process;
- communication of the AI system’s capabilities and limitations;
- need for AI systems to be identifiable as such, ensuring that users know they are interacting with an AI system.
The guidelines are a first step taken by the EU to ensure that AI is more accessible to citizens. This appears even more vital given the integral role AI technology plays in our daily lives.
The Finnish success story
Citizen-centred AI is also being shaped by Pekka Ala-Pietilä, the Finnish chair of the European Commission’s High-Level Expert Group on Artificial Intelligence. His home country, at the forefront of the platform economy, sees the development of technology as being intertwined with democracy: the more the citizens participate, the more will AI systems benefit society at large. The Finnish AI political steering group actively encourages citizens and stakeholders to participate in the policymaking process, with the dedicated hashtags #tekoälyaika and #aiera. Also, both the Commission’s High-Level Expert Group on Artificial Intelligence and its forum the European AI Alliance follow a multi-stakeholder approach for establishing a broad, open and inclusive discussion for the development of AI.
But how should citizens with a non-technical background contribute to developing technological innovation? The Finnish answer is Elements of AI. This educational online course (massive open on-line course or MOOC) teaching AI basics is a success story: within one year, over 160 000 people registered. The course is available in English, Finnish and Swedish. Due to its massive success, similar elementary AI courses have been developed in the Netherlands with ‘De Nationale AI Cursus’ ai-cursus.nl in Dutch and in Brazil with ‘Inteligência Artificial Fundamentos’ in Portuguese.
Teemu Roos, founder of Elements of AI, strongly believes in the power of society and aims to make his course go global. He is currently building up international partnerships in other industrial countries as well as in Africa. Roos created Elements of AI intending to design an open, easily accessible course in which every citizen should feel encouraged to participate. In this respect, Elements of AI reflects fundamental European values such as equality, human rights and democracy.
Teemu Roos : How well can a society adapt to the changes brought by AI? The most important questions are not really technological – they are political. In order to democratically shape these decisions, wide-ranging public discussions are needed. If some populations are poorly represented, this discussion is not possible.
Pooling European AI research
Until recently, the AI scientific community in Europe has been fragmented. In order to benefit from the synergies, researchers established CLAIRE, the pan-European Confederation of Laboratories for Artificial Intelligence Research in Europe. This network pools over 14 000 researchers from the wider Europe. While the main objective is to strengthen AI research and innovation, CLAIRE ensures European-wide regional representation within informal advisory groups. Experts play a key role when it comes to regulating AI. As CLAIRE co-founder Holger Hoos puts it:
Holger H. Hoos : Citizens as well as decision-makers, politicians and administrative staff are generally not sufficiently educated about AI. However, every political decision-making body depends on expertise, most importantly because AI has far-ranging societal impacts. Unfortunately, humans tend to overestimate what technology can do for us. It would be highly problematic if less AI-knowledgeable people were to develop, deploy, monitor and maintain AI systems.
What is meant by ‘understanding AI’?
Nevertheless, the debate about AI could benefit from wider understanding of the technology enabling it – namely algorithmic programmes and machine learning. Reports and articles often emphasise knowledge and digital skills – but does this entail knowing how an algorithm is programmed? What exactly is meant by ‘understanding AI’? One option would be to differentiate between:
- transparency of the algorithms (on a technical level);
- analytical-educational skills (using AI systems for work or leisure purposes);
- consciousness and public acceptance (awareness about AI systems in operation).
This distinction could help in interlinking ethical as well as technological challenges, as the majority of AI development entails both aspects. Social acceptance therefore requires a meaningful dialogue about defining ‘human-centred AI’, in relation to daily use and differing levels of understanding, for everyone.
Citizen-centred AI communication
As Finland takes over the EU presidency in July 2019, many expect AI to be one of the mandate’s priorities. Teemu Roos acknowledges the opportunity to place AI at the core of Europe’s future – both for policymakers and citizens:
Teemu Roos : The EU has a central role as AI regulator: development of AI technology by private corporations or regimes with limited interest in upholding human rights fundamentally challenges democracy. Therefore, powerful intergovernmental actors are essential. The Finnish EU Presidency is an opportunity to share the Finnish success stories in education and human-centric technology.
Finland seems well prepared for the all-encompassing changes AI is bringing about. This is also due to the success of Elements of AI. The project surpassed by far its goal to educate 1% of the Finnish population in AI. Therefore, ‘normalising’ AI for every European would ensure the use of smart and at the same time trusted technology across all sectors. Societal acceptance of AI requires, first of all, meaningful dialogue about the ethical and technological challenges facing everyone in the EU.
• A scientist’s opinion : Interview with Teemu Roos about AI education
• A scientist’s opinion : Interview with Holger H. Hoos about AI education
• EU Project : AI4EU
• EU Project : FANDANGO
• EU Project : DYMO
• EU Project : EARTIL