The EU's approaches to human-centric artificial intelligence reached thousands of international audiences onsite or online at the Paris Peace Forum on 11-12 November 2022 and at the largest climate gathering of the year COP27 on 14-15 Nov 2022 in Sharm El Sheikh.

European AI at PPF

European Commission Director of Artificial Intelligence and Digital Industry Lucilla Sioli was a speaker at the Roundtable on “Managing AI Risk: Perspectives from around the world”, on 11th November 2022. 

“The EU is the only organisation at this moment that has proposed a comprehensive framework for high-risk AI applications. And this is not about regulating AI as a technology, it’s about regulating the development of certain AI applications in certain contexts at the very beginning, said Director Sioli at the Roundtable co-organized with the OECD.AI. 

She said it depends on how you apply your technology and in which context you are talking. “So, what we have done is we have introduced a number of rules because we want to be able to make sure that AI respects the fundamental rights and safety standards that we have achieved over the years,” Director Sioli said. She added that for the EU this piece of regulation is not introducing new rights but is a mechanism to make sure that the applications of certain AI systems really respect the EU's fundamental rights.

The EU has identified some applications that it considered to be high risk based on evidence it found and what it has introduced is mechanisms to make sure that before these applications are put on the market, they are checked from certain points of view including data quality, robustness and accuracy, the information they provide to the users or human oversights they introduce.

This is a system which the EU has followed for safety kinds of applications. These are checks that require the developers to make sure before they put their systems on the market, and market surveillance organisations are available to step up if there are incidents afterwards.

For the EU to facilitate enterprises and companies that develop AI and ensure that the requirements are checked, the EU is working on standards because a lot of this is going to be self-assessment by companies and therefore the availability of standards is very important in the EU, Director Sioli said.

She explained that companies can show they have followed an agreed set of standards then this is sufficient to show compliance with the requirements that the EU has.

Therefore, the EU is working on the standardization part now and “we are about to issue a mandate to the EU standardisation organisations and so we work with many different stakeholders,” Director Sioli said. She added that the EU is working within the context of the trade and technology councils so it needs to have a catalogue of tools, measurements, of instruments that it can use and that will have an impact on the standards that it develops. This collaborative approach to the regulation of AI will make life easier for companies in other parts of the world.

“This is the EU approach and I think it’s very much in line with the UNESCO recommendations and with the works that have been done by the OECD and other international organisations,” Director Sioli said.

Also at the Roundtable on “Managing AI Risk: Perspectives from around the world”, Dr Gry Hasselbalch said “we have to consider how these [AI] systems have evolved over time as part of a general socio-technical development… Certain interests have been shaping the power dynamics of the negotiations on how we transform regional and local systems into larger global systems.”

In her expert opinions from the InTouchAI.eu project, the most important part is to consider how different interests are prioritized in for example risk mitigation frameworks. “Which risks are we considering when we are making risk mitigation initiatives; something like the EU’s AI Act or the UNESCO recommendation are actually putting a threshold of our risk tolerance.”

And that is the EU human-centric approach, Hasselbalch said adding that she also wants to see the planet-centric approach because it’s including a systemic understanding of the role of human beings in our environments. From that perspective, the key risk would always be the risk to humans and the risk to the environment where she saw different AI initiatives coming together.

In the Roundtable Discussion “Making AI work for international development” on 11th November 2022 Dr Gianluca Misuraca represented InTouchAI.eu to introduce the project and its objectives in bringing a holistic human-centric AI governance approach that honours universal values, fundamental rights, and principles to the global partners.

“Sometimes we forget that technologies should serve humans, not the other way around. That’s why this human-centric approach is being promoted,” said Dr Misuraca, explaining that this is embodied in the values and the principles of the EU.

IntouchAI.eu is set up to promote these approaches and the project was selected as one of the 60 projects to be showcased at the PPF.

“We are trying to promote the activities that are often associated with the AI Act, but the strategy of the EU is not only the AI Act, which is an important piece. There’s a plan of action that all the member states have signed and has been revised in 2021 as a part of the AI package.”

The project is one of the activities and it focuses on the international dimension of AI. “The EU is now partnering with organisations and other like-minded countries all over the world. What we have been doing so far and what we do is to support the EU and commission colleagues in policy dialogues, for instance with different the TTC with the US or with others like India, Japan and others.”

InTouchAI.eu also does research and gathers intelligence to have mutual understanding and for the same ground in discussions on what the EU means by human-centric AI. The roles of human rights or how we protect data in one way or the other are important when it comes to the discussion with the OECD, UNESCO and others by providing some evidence through analysing practices and examples in the EU and across the world.

Dr Misuraca said, “What is also important for us is the technology diplomacy dimension of this activity” adding that the civil society, the activists, and the public also have roles to play and that’s why multi-track diplomacy is needed in this field. Therefore, the project’s activities are multilateralism and multi-stakeholders, supporting the official debates and influencing the discussion in the public space in the EU and worldwide. Dr Misuraca said that the future of this initiative may be even enlarged to other regions in the world to find compromises and solutions, that often come out from specific projects and initiatives, like InTouchAI.eu.

In the two-day PPF, InTouchAI.eu also had a booth where experts shared the project’s activities, the recent developments of AI regulations and the AI landscape in the EU. The experts highlighted the EU's efforts to promote the development of an AI that is safe, puts people and their livelihoods at the centre and promote human rights along with internationally recognised values. 

With the strong participation of European AI at the PPF by the leaders of a few directorates including the European Commission’s Service for Foreign Policy Instruments (FPI) and the Directorate General for Communications Networks, Content and Technology (DG Connect) FPI and experts’ contributions and the above activities, European AI added to the success of the 5th PPF. The PPF analysed ways to mitigate multiple shocks and the socio-economic consequences of crises while supporting multilateralism and a more cooperative world.

European AI at COP27

The European approaches to AI continued its presence and extended the messages to COP27 where InTouchAI.eu experts joined two workshops: An EU Approach to AI for Sustainability at the Danish Pavilion and the Re-Imagining Common Goods in the Age of AI at the UNFCCC Pavilion.

In the workshop co-organised with the Danish Pavilion, InTouchAI.eu Project Leader Dr Misuraca moderated a panel of experts coming from policymaking and academics on the relevance of the proposed EU AI Act and its impact on fostering the adoption of sustainable AI in Europe and beyond, together with the implementation of the European Green Deal.

The discussion centred an argument that AI and digital technologies can play an important role in achieving climate neutrality goals, restoring the planet’s biodiversity, and limiting pollution in the coming decades, however, the promise of the benefits of AI for the green transition is also tested against its risks and transformational effects on society.

This is where Europe’s approach to AI, which is based on the twin objective of excellence and trust, becomes relevant because its objective is to ensure AI innovation with socio-economic, and environmental well-being and people’s safety and fundamental rights at the centre. And realising this vision requires coordination in the context of international organisations and possibly direct collaboration with like-minded partners.

Policy experts discussed the Coordinated Plan of Action agreed by EU member states in April 2021 which included a dedicated section titled “Bring AI into play for climate and environment”, and this aims at supporting the EU objectives to cut greenhouse gas emissions by at least 55 % by 2030 and to be climate neutral by 2050.

Realising this vision requires coordination in the context of international organisations and possibly direct collaboration with like-minded partners. The EU will indeed step up its bilateral and two multilateral efforts to support the establishment of a global level playing field for the trustworthy and ethical use of AI.

While acknowledging the potential of AI solutions and technology in supporting the EU to meet climate goals, experts also call for stronger attention toward the sustainable use of AI as tools, considering the impacts of AI on social and environmental sustainability.

The EU’s effort in ensuring the sustainability of AI and minimizing its negative impacts was demonstrated by research done at the Bonn Sustainable AI Lab, which InTouchAI.eu Research Lead Dr Gry Hasselbalch initiated and collaborated with a few joined research and dissemination activities, or through a European best practice on how the European energy industry can contribute to AI for sustainability.

The session at the Danish Pavilion contributed to the ongoing discussion and promotion of the EU’s approach to AI by building on the insights gained from previous expert consultations and research on AI for Sustainability within the framework of InTouchAI.eu.

New views and recommendations have been discussed and elaborated on the role of the EU in promoting sustainable AI worldwide. The panel was moderated by Gianluca Misuraca, InTouchAI.eu Team Leader and Senior Expert in Technology Diplomacy and featured five European experts, including InTouchAI.eu Key Expert and Research Lead Gry Hasselbalch, representing the European twin objective of excellence and trust both in academia and industry.

European Approach to Artificial Intelligence for Sustainability introduced that AI technologies have enormous potential in fighting the climate crisis, however, they also have negative environmental impacts, which the EU is considering when developing its human-centric approach to these technologies and their implications. 

This panel leveraged the insights and research suggestions gathered at the workshop that InTouchAI.eu organised at the World EXPO in Dubai in March and focused also on how the European approach to AI is perceived and influential internationally. 

From the discussion at EXPO Dubai emerged key recommendations. Starting from those recommendations, this panel’s main objective was to discuss the relevance of the proposed EU AI Act and its impact on fostering the adoption of sustainable AI in Europe and beyond, together with the implementation of the European Green Deal. 

This panel fitted into a broader international discussion, for instance, the session “What is the environmental footprint of AI” organised by the OECD and our partner event “Re-Imagining Common Goods in the Age of AI” organised by the Bonn Sustainable AI Lab at the UNFCCC Pavilion, which highlighted the need for a socio-political perspective on AI issues when considering the impacts of AI on social and environmental sustainability.

The discussions continued broader perspectives at the Re-Imagining Common Goods in the Age of AI workshop at the UNFCCC where the project experts and other speakers emphasized the inclusivity, the larger perspectives of common good infrastructure including the digital systems, and a global united effort needed for safe and effective use of AI systems.

The speakers agreed that the EU human-centric approach to AI has many advanced concepts, and it needs a stronger consideration of the national, social, and environmental impacts of AI. “We need to join voices and forces and we need to act with a sense of urgency,” said Dr Misuraca, adding that We still have time to develop human-centric AI and make it right at the beginning of this AI age.

Dr Gry Hasselbalch stressed the need to consider the issue of data pollution in the context of the development of AI technologies. Data pollution is an environmental problem of the big data age we are currently living.

She stressed that the recent political focus on AI, trustworthiness, and sustainability we have to see in the context of a broader global policy agenda regarding the environmental impact of science and technology which has evolved over decades. Geo-political concerns with the adverse effects and disturbance of the natural environment and ecosystems caused by manmade technological progress have always been at the core of the global sustainable development agenda.

The first global Human Environment Conference that took place in Stockholm, Sweden in 1972 recognised the impact of human science and technology on the environment, stating the urgency to collaborate and act globally stressing among others the need to develop “Environmentally Sound Technologies” (ESTs) to help solve environmental problems, while in and by themselves being environmentally friendly.

Dr Hasselbalch pointed out that in recent years, the EU has been leading the development of policies on AI that include reflections on the technology’s social, ethical and environmental impact and its approach thus representing a best practice with the concepts being included in policies and regulations starting with the GDPR, and continuing with the Coordinated Plan on AI, the European Green Deal that stresses that ‘sustainability’ must be a core point of departure for the development of not only AI technologies but a digitised society in general. and then the AI Act itself that more specifically addresses the impacts on our social and personal environment.

Many other sessions at COP27 or alongside this global event have also been focusing on the environmental impact of AI technologies and the implications of using AI as a lever for fighting climate change and making a greener society. 

InTouch.AI contributed to this with the panel discussion “A European Approach to Artificial Intelligence for Sustainability” and the partner event “ReImagining Common Goods in the Age of AI” organised by the Bonn Sustainable AI Lab, that Dr Yardımcı Geyikçi moderated and to which Dr Misuraca and Research Co-Lead Dr Hasselbalch participated to as speakers.

PPF

Managing AI risk: perspectives from around the world: 1.3K views

https://www.youtube.com/watch?v=leFUxSTG1pI&list=PLhHMwTfbaJ0TrUDQA-fey…

AI in International Development: 836 views

https://www.youtube.com/watch?v=8yaz4zK5Gqk&list=PLhHMwTfbaJ0TrUDQA-fey…

COP27

An EU Approach to Artificial Intelligence for Sustainability => 50+ onsite view??

Re-Imagining Common Goods in the Age of AI, COP 27 => 90 views + onsite???

https://www.youtube.com/watch?v=K9zmt_UI6iA&list=PLBcZ22cUY9RJxaTQ6OnqLAvXQy9FykHiy&index=37