On the 12th of December 2022, the IEAI and TUM Think Tank hosted an in-person panel discussion focused on the diverse “Global Insights on AI Governance.” Professor Urs Gasser, from the Technical University of Munich, served as moderator to the speakers: Professor Laetitia Onyejegbu from the University of Port-Harcourt, Dr. Elisabeth Sylvan, Dr. Armando Guio Espanol from the Berkman Klein Center and Professor Christoph Lütge, who discussed the challenges to creating global approaches to AI governance that incorporate cultural distinctions and diversity in values and priorities.
Professor Onyejegbu began the discussion by presenting her research into how Artificial Intelligence and Machine Learning can increase food security and crop production. This research is motivated by the significant food insecurity in Africa caused by climate change, whether it is a pest infestation or poor weather conditions. She explained how her team studies the Internet of Things and AI for crop prediction through mounted frames that monitor factors such as soil fertility or crop disease. By doing so, an AI model collects and trains itself with farm data from the concerned areas, which is then stored on an online database where farmers can have end access to what is happening on their grounds. This way, farmers can increase crop yield by locating current and potential problems such as treating an infestation, which fertilizer to use, etc. Although promising, a challenge is financial: farmers might not have enough funds to buy this system. Therefore, suppliers and innovators must place efforts to limit the costs of this technology and for the governments to take action to support and provide farmers with these systems. This opened discussion for the necessity of enablers and stakeholders within the government to enforce these technological developments and frameworks.
The floor was handed to Dr. Sylvan, who delved into the importance of AI-based tech in education and the trends occurring in the classrooms regarding advanced technology. This topic became relevant due to the surge of COVID-19, which provoked drastic changes in the classroom. She explained how academic institutions switching to online learning was uneven throughout the United States, as each state or municipality had an individual plan. Her research found that traditional learning was challenged, and technology within and outside the classroom changed drastically. For example, hybrid learning became more habitual to mitigate potential interruptions throughout the year. A challenge from this is that most tools necessary to facilitate online learning are not publicly funded; therefore, not all institutions can have access, which can trigger detrimental social cleavages.
The data that these tools collect foster the potential for improving learning methods and creating new developments toward creating a Blueprint for an AI bill of rights. However, this is challenged by the need for adequate funding, and there remains work needed in the education of AI for administrations, teachers, or students.
This evolution of AI within education offers the opportunity to create personalized learning environments for individual subjects or students and overall life. In the future, it can change the material assessment methods, and the collected data can offer more knowledge to educators on how to teach more effectively. However, these possibilities open the debate on who can access these databases and what safeguards must be used to ensure that it is collected and used ethically.
Professor Espanol followed with an introduction to his work researching and assisting governments in Latin America in developing AI policies. He has worked with the OECD and particularly concentrated on creating ethical AI practices in Columbia. When asked about the government’s role when navigating the balance between technological developments and potential implications, his research found that a multistakeholder approach was the most fruitful when designing AI strategies. In the case of information asymmetries, due to a lack of knowledge of data and algorithms by most governmental actors, he found it was essential to create spaces with experts in various fields to convey various tools and information to create effective AI strategies. In the case of Columbia, Professor Espanol and his team created guidance on AI ethics, digital skills, governance, and improving information systems. An encountered issue once again is the lack of adequate funding. Finally, he elaborated on the importance of collaboration between experts globally through policy labs by bringing various stakeholders together to work on policies.
Professor Onyejegbu further elaborated on the topic of the capacity problem and knowledge about AI being unequally distributed, as she explained the role of university and STEM education within schooling from an educator’s perspective. She believes it is crucial to begin from the grassroots to create awareness of AI, as it allows people to become more familiar with and build trust with the concept. At the same time, governments must place more importance on funding AI projects to aid people who have the capabilities to create these AI technologies, as this is one of the starting motivators to foster the creation of AI.
Dr. Sylvan also chimed in on the capacity and inclusionary problem, as frequently, young people are not part of the conversation of the future of AI. She also agreed on further training, support, and funding necessary to mitigate this problem. She spoke on a program she is currently working on, which focuses on supporting the global network of AI, hosting learners worldwide to participate and connect with expert researchers, and working on projects that foster connection and learning. The hope is that these projects are built on a broader level and can become interconnected with themselves and powerful institutions. These platforms provide the possibility to have a conversation which aids in taking steps forward and effecting change.
Following this, Professor Gasser brought in Professor Lütge by asking about IEAI and how it developed a “multiple-instutionary” program. Professor Lütge then explained that this project was launched in 2021 in collaboration with NYU and offered an open AI ethics course system with numerous online modules accessible online to everyone. Another initiative was GAIEC, created during the COVID-19 pandemic to bring people and institutions together around the globe by creating a network to submit proposals or foster knowledge.
Although there has been much progress in AI, we remain in the early stages of its development and the role it plays in various subjects in our lives. Like past technological innovations, we need time to evaluate how AI will establish itself in our lives and shape our future. For this reason, it points to the question of informational asymmetry existing despite the collaborative efforts and how to help policymakers decide on their uncertainty around the subject. Professor Espanol presents the concept of AI Sandboxes which he has found in his personal experience to be very fruitful; in this context, it allows researchers and governments to have an insight into how AI technologies and policies may interact in society. With this, we can take steps towards action, whether creating more knowledge, acceptance, or regulation of risks. He believes these spaces also push towards more transparency and well-rounded policymaking.
Professor Gasser then questions the proper resources to gain more information on AI Regulation and asks Professor Onyejegbu about the intellectual influences on policymaking, specifically in the case of Nigeria. In her work, she observed that when it comes to the AI policy in Nigeria, they focused on the contributions and publications of International Organizations such as UNESCO for inspiration to alter it to fit Nigerian society
This panel ended with a discussion between the panelists and the audience about their research projects and how AI global governance could evolve in the future. The IEAI was proud to work with the TUM Think Tank to create an event delving into the importance of AI Governance, its interactions on a global scale, and what steps could be made to optimize its evolution in the future.