.png)
.png){kind=small}
Stefan Hofmann, Curtin University
Going to Geneva earlier this year felt like going to the centre of the world. I’ve never experienced being surrounded by such a diverse group of people, all contributing ideas to address the world’s current and future challenges. For this reason, the city is often called the centre of international diplomacy, given that it is home to the largest number of international organisations in the world. I was able to hear from some of these diverse voices at the AI Governance Day as part of the 2024 AI for Good Global Summit, where we heard from tech companies, UN officials, and global world leaders on how to tackle the governance of artificial intelligence.
After listening to leaders from both developed and developing countries at the AI Governance Day, a fascinating divide had emerged. World leaders from developed countries, commonly called the Global North, spoke about AI with an undertone of fear. Heavy emphasis was placed on highlighting issues such as algorithmic bias, explainable AI, and copyright issues. However, I couldn't help but notice the change of tone when leaders from the Global South spoke. An undertone of fear was replaced by hope, and an urgency to tackle problems was replaced by an urgency to continue rolling out the technology. Emma Theofelus, Minister of IT in Namibia, dismissed the need for an AI Safety Institute in Namibia as the name implies that AI is a threat to the country, demonstrating her optimism toward the technology.
This conclusion gathered from listening to various leaders at the AI Governance Day is not just a feeling but a phenomenon observed in global research. A study by Ipsos conducted in 2023 found that the US, New Zealand, Canada, Great Britain, and Australia were the five most nervous countries surrounding the use of AI in products and services. On the other hand, Turkey, Mexico, Indonesia, South Korea, and Thailand were most excited about products and services incorporating artificial intelligence. (1)
It's not hard to see why this is the case. People from countries part of the global north, such as Australia, who fear that increasing adoption of AI will create problems such as triggering mass job losses, create excuses to stop funding government services like education, and cause complex copyright issues, will most likely have more to lose than others in the global south (2). This is in contrast to people in the Global South who can see the incredible opportunities that AI can bring to their country. In Geneva, we heard from a researcher at Google who discussed his work with Med-PaLM, and how language models trained on some of the best health specialists in the world can bring their knowledge to the rest of the world in the form of a chatbot. Other people also discussed how AI can drop the cost of first-level cervical cancer screenings from $700 to only a couple dollars, and provide better predictions on the supply-chain of vaccines in Tanzania. As a global shift away from rural areas into cities continues, AI can help close the resourcing gap by automating farm vehicles and provide insights to improve crop yields. AI is also analysing post-earthquake reconstruction needs in countries like Nepal (3). With case studies like these, it's not hard to see why some countries are more willing than others to look past discussions of safety, when AI can provide mechanisms to solve pressing issues within their economy.
Unfortunately the solution is not as simple as providing health, farming, and education industries in these countries with off the shelf AI solutions. The most important ingredient for creating an artificial intelligence algorithm is data, and this data, with all of its intricacies and biases, directly dictates how effective the AI algorithm will be. At the AI Governance Day, the Director of the International Labour Organisation, Gilbert Houngbo, discussed the unfortunate reality created when the majority of data gathering and development of AI occurs in the global north, but the greatest need for these technologies is in the global south. Houngbo, and many others at the conference stressed that successful implementation of AI solutions in these countries requires robust, on the ground data collection methods so that AI models can reflect local knowledge and geographic variances without reinforcing biases that foreign data may bring. The importance of sound data collection methodologies is seen in case studies such as Amazon's AI Recruitment tool, which analysed successful job applications over the last ten years to be able to automate selection processes. Unfortunately, as most job applicants were men over this period, the algorithm automatically dismissed many female candidates (4). The problems that a lack of diversity in data collection brings was also emphasised in 2015 when Google Photos was found to be labelling pictures of dark-skinned people as 'Gorillas' (5).
Despite these challenges, world leaders representing the global south at the AI Governance Day were eager to highlight their investments into their AI futures. Dr. Mercedes Aramendía highlighted at the governance day how Uruguay set a goal to train 80 000 students in AI development in 2023, and Mauricio Lizcano, the Columbian IT and Communications minister also showcased their investment in setting up AI bootcamps to advance local understanding. Investment in local AI talent enables these countries to harness the full potential of AI in being able to advance their economies, while also ensuring that they are fit for purpose within their local environments. With populations that are also statistically more likely to embrace AI solutions, and investment in onshore AI development capabilities, the global south has great potential to harness the capabilities of AI in solving some of their most pressing issues.
-------
The views and opinions expressed by Global Voices Fellows do not necessarily reflect those of the organisation or its staff.