Reflections and Key Takeaways from GPAI Summit 2024
The Global Partnership on AI (GPAI) Summit, held at the Palace of Serbia in Belgrade, Serbia, on December 3–4, 2024, was a testament to the global collaboration in shaping the future impacts of artificial intelligence (AI). Delegations from 44 countries, policymakers, academics, and technology leaders worldwide came together. I had the privilege to witness essential discussions on how AI can be a transformative force for humanity. I would like to share with you what I learned here, the most remarkable talks for me, and the notes I took during the GPAI Experts session I had the chance to attend. You can access the details of the summit here.
Day One: A Strong Start for Human-Centred AI
I can only say this about the atmosphere as I entered the summit. Everyone, from the top decision-makers to AI experts worldwide, seemed ready to discuss AI’s enormous opportunities and risks. I could see this spirit clearly in the opening remarks.
OECD Secretary-General Mathias Cormann did not shy away from emphasizing the role of AI in economic growth and sustainable development and the potential for digital inequalities. Again, this powerful statement made the point that technology is an opportunity and a responsibility that must be managed carefully and will only affect some equally globally.
Serbian Prime Minister Miloš Vučević and Serbian National Assembly President Ana Brnabić emphasized the importance of integrating AI into public services for Serbia. They explained how AI is being used to ensure equal opportunities in education, especially in rural areas, saying: “AI has enabled us to provide the same quality of education from big cities to the most remote rural areas. It is not just a technology but the key to equal opportunities.”
Dr. Amandeep Singh Gill, the United Nations Secretary-General’s Representative for Technology, emphasized in his speech the transformative effects of AI on societies and the urgent need for ethical AI regulation. He drew attention to the opportunities provided by technology from the early days of the internet to the present and the social and ethical problems that AI can create if misused. Stating that the digital divide can deepen due to a lack of education and infrastructure, especially in developing countries, Gill stressed the importance of fair Regulation with the words, “If we cannot make the benefits of AI accessible to everyone, the digital divide will widen, and this will create not only a technological but also a social crisis.” He drew attention to the need for a fair distribution of the great opportunities offered by AI in areas such as agriculture, health, and climate change. He defended the need for a scientific assessment platform at the global level while addressing the harm that disinformation and its social effects have on societies. Gill concluded his speech by reminding us that understanding, regulating, and directing the future of AI is our common duty.
Took a Deep Look at AI Safety with Stuart Russell
Stuart Russell, a professor of computer science at the University of California, Berkeley, is one of the most respected people in the world regarding AI. Listening to his impressive speech on “Trustworthy AI” inspired me. Russell drew attention to the serious shortcomings of current approaches to the security of AI systems. He emphasized the need for new strategies for this technology to remain under human control in the future. In his speech, he said, “An AI system that is not under human control will sooner or later find its vulnerabilities and has the potential to cause harm. Therefore, we need formal assurance mechanisms to ensure that the systems we develop benefit people and are safe.” With this statement, Russell drew attention to the need for a radical paradigm shift in AI design and reminded us that security is a technical necessity and an ethical responsibility.
He stated that the development of AI should become a discipline with high safety standards, like aviation or nuclear energy. “Patching misaligned systems to make them secure is not sustainable,” he said of current approaches, noting that current methods will not work in the long term. Instead, he argued that algorithms should optimize for specific goals and ensure that these goals are aligned with human values. “The potential benefits of AI are enormous; we can advance humanity in every area, from education to health. But we can only achieve these benefits with systems that are developed safely and ethically,” Russell said, emphasizing the urgency and importance of this transformation. His speech powerfully illustrated the need for a systematic and scientific approach so that AI can become an opportunity rather than a threat to humanity.
The panel discussed the role of governments in driving global AI innovation under the title ‘Unlocking AI Potential: The Role of Governments in Global Innovation.’ Participants focused on AI’s potential to boost economic growth, open new markets, and enable inclusive, sustainable economic progress. The panel also emphasized the crucial role of international cooperation in AI innovation. They discussed how strategic government initiatives, coupled with collaboration between countries, can harness AI’s transformative power and ensure that the benefits of AI are accessible to all.
Panel: Unlocking the Potential of AI — The State’s Role in Global Innovation
- Jelena Begović (Serbian Minister of Science and Technology Development): Begović emphasized the importance of governments increasing their investments in science and R&D and noted that Serbia prioritized technology education by adding compulsory programming courses to the curriculum in 2017. She cited Serbia’s investments in physical and digital infrastructures, such as data centers and the Leonardo Supercomputer Center in Bologna, stating that infrastructure is a key element for AI innovation.
- Clara Chappaz (French Secretary of State for AI and Digital): Chappaz detailed France’s National Strategy for AI, which was established in 2018. She noted that this strategy aims to make technology accessible to all, sustainable, and inclusive. She drew attention to education and awareness-raising and stated that governments are responsible for reducing concerns about AI.
- Jitin Prasada (Minister of Electronics and Information Technology, India): Prasada spoke about India’s efforts to empower its large population with digitalization and AI technologies. He emphasized that they have created an ecosystem based on public-private partnerships and how AI can bridge access gaps through language models. He also stresses the importance of balancing innovation with data privacy and protecting citizens’ rights.
- Ulrik Vestergaard Knudsen (OECD Deputy Secretary-General): Knudsen said governments must invest in R&D to support AI’s inclusive and reliable development. He also emphasized the importance of international cooperation and regulatory systems’ interoperability. “Establishing a globally harmonized AI regulatory system can be challenging, but agreeing on minimum standards is vital,” he said. He cited the OECD AI Principles as a guide in this regard. Ulrik noted that he was not sure that the EU AI Law was smartly prepared, but he also pointed out that the definition of “AI System” was smart enough and that it was the OECD’s proposal. Although some did not accept this interesting statement, it continued to be discussed in later sessions.
Yoshua Bengio: Risks and Opportunities of AI
Yoshua Bengio, founder and scientific director of the Mila—Quebec AI Institute, evaluated AI's technological and algorithmic developments in recent years, emphasizing that these advances carry great benefits and severe threats to society.
Discussing the potential benefits of AI, he said, “AI can play an important role in solving our most pressing global problems, such as the climate crisis,” drawing attention to the technology’s positive effects. However, he also emphasized the risk of misuse of this powerful technology.
He stated that while AI’s capabilities in cybersecurity can be useful, they can become a dangerous weapon in the hands of bad actors. He warned that AI can be used to discover new drugs and design deadly biological weapons, especially in areas such as biochemistry.
Bengio, addressing the risks of AI’s advancement, stated that societies and states must act quickly enough to understand and manage these risks. “Waiting for these risks to materialize is not an option because some potential outcomes could be catastrophic,” he said, drawing attention to the importance of international cooperation and Regulation. Bengio noted that an international scientific report to be presented in Paris in February is being worked on. He said this report will provide a comprehensive framework for the security of advanced AI systems, their potential risks, and the precautions that can be taken against them.
Finally, he outlined a roadmap for future regulations on AI. Emphasizing that ethics and transparency principles should be taken as the basis, he said, “We must create a regulatory system that rewards ethical behavior among companies.” He also concluded his speech by stating that allocating more resources to technical research will be critical in developing security and control mechanisms.
Day Two: Sustainable Development and the Place of AI in Society
The second day of the summit continued with the discussion of AI in a broader societal context. Both promising developments and challenges were discussed in these sessions. On the other hand, the second was even more exciting for me. The report launch of the project we carried out with GPAI, the International Center of Expertise in Montreal on AI — CEIMIA, and Oxford Insight for 3 months last summer was held today. Since the owner of the relevant project in Türkiye is the Digital Transformation Office Presidency. As the project coordinator, I had the chance to meet closely with the appropriate stakeholders at the summit and consult on future work. They also invited me to the GPAI Experts meeting session, which was a closed session, which you can find at the end of my article!
Vladimir Crnojević: Artificial Intelligence for Sustainability
While discussing the role of AI in the context of the United Nations Sustainable Development Goals (SDGs), Vladimir Crnojević from the BioSense Institute said, “AI can reduce inequalities in health and education, but at the same time, we need to be more careful about energy consumption and environmental impacts.” This warning highlighted technology’s environmental costs and contribution to sustainable development. I highly recommend you look at this article on this subject in Nature.
AI Regulation: What have we learned so far?
In his speech at the GPAI Summit, Gabriele Mazzini, one of the leading figures in AI regulation at the European Commission, emphasized the need for a comprehensive and harmonized regulatory framework to manage AI technology’s impact on human life responsibly.
Focusing on the EU’s AI Act, Mazzini stated that this law aims to be a reference point for the EU and globally. “We must spread the benefits of AI throughout society while minimizing its risks,” Mazzini said, adding that the law encourages safe and ethical use of AI. Noting that the regulatory framework should support security and transparency standards without restricting innovation, Mazzini said: “Regulation should be a guide, not an obstacle. We must balance protecting fundamental rights and democratic values while encouraging innovation.” He also emphasized the importance of international cooperation in AI-related regulations and stated that regulations should be made interoperable. “Different regulatory approaches around the world need to converge on minimum standards,” he said.
Finally, Mazzini concluded his speech by noting that the future impacts of AI could reach unpredictable dimensions, saying, “Taking precautions is not only a responsibility but also a way to ensure that technology benefits humanity.” The session continued with an in-depth discussion by Merve Hickok, Maja Bogataj Jancic, Kyoto Yoshinaga, Till Klein, and Tuleushin Kanysh on understanding and navigating AI regulations’ technological and societal impacts.
Bertrand Monthubert: The Power of Public Data and Türkiye’s PSDS Project
One of the highlights of the second day was the panel hosted by Bertrand Monthubert, Co-Chair of the GPAI Data Governance Working Group.
Monthubert discussed how public sector data can create value for AI and launched the “The Role of Government as a Provider of Data for AI” report conducted jointly by GPAI, CEIMIA, and Oxford Insight. Türkiye’s Public Sector Data Space (PSDS) project was one of the three countries covered as an application example in this report. It was cited as an example by Monrhubert at the launch: “Türkiye’s work in this area can guide other countries on how public data can be used effectively.” With this comment, of course, I experienced the proudest moment of the summit. The fact that this work, which I have been involved in since the beginning, sets an example at the international level has satisfied me both in terms of my country and my vision. If you hear about this project for the first time, you can review its details here!
Details of the PSDS Project:
- The project is being carried out in cooperation with TÜİK and TÜBİTAK BİLGEM, with the Presidency Digital Transformation Office coordinating it.
- Innovative solutions, such as anonymizing occupational health and safety datasets and making them suitable for AI applications, were presented to the Ministry of Labor and Social Security as a pilot application.
- The Role of Government as a Provider of Data for AI report emphasized PSDS’s potential to transform public services while protecting citizen privacy.
Monthubert’s assessment introduced Türkiye’s innovative approaches to AI and data governance at a global level.
I had the chance to consult with Stefan Janusz from CEIMIA one-on-one and talk with him about other work they were doing at CEIMIA. I can say that the summit became real for me towards the end of the second day, as he introduced me to Bertrand Monthubert, which was the reason I was invited to the GPAI Experts session. So I have to thank Stefan just specially for that.
GPAI Uzmanları Session
This session was coordinated by Karine Perset, Chair of the OECD AI Governance Working Group, where GPAI Experts from around the world presented their current work, and others asked questions to representatives from that country. Although I am not an official GPAI Expert, I had the chance to attend this meeting because of the work we have been doing together recently. The meeting, which was attended by Bertrand Monthubert, Chair of the GPAI Data Governance Working Group that I mentioned earlier, and many countries, was very valuable for me to have the chance to share some of the information we mentioned in the report on the Public Sector Data Space. I think I have decided once again that some events should be held face-to-face :)
Conclusion: Shaping the Future with a Unified Vision
The GPAI Summit in Belgrade discussed technology and expressed a shared vision for social transformation and a better future for humanity. The solutions offered by AI in critical areas such as health, education, environment, and public services were eye-opening and exciting. However, there was a consensus that this technology should be managed responsibly, transparently, and ethically. It paved the way for valuable discussions on issues without consensus, such as the EU AI Law, which must be rethought. I saw at this summit that AI can create significant value for humanity as a tool. Still, this value needs to be managed carefully and addressed with a sense of shared responsibility. The fact that Türkiye is leading this transformation with projects like PSDS and that it is a remarkable success in the international arena is definitely a note that I should add here. In addition, Belgrade, like other Balkan cities, was familiar with its coffee and pastries.
