UN AI for Good Summit Explores Risks and Opportunities of Generative AI

2023-07-10

On July 6th and 7th, the United Nations hosted the 6th AI for Good Global Summit. In the panel discussion titled "The Next Wave of AI for Good - Towards 2030," experts in generative AI highlighted the risks associated with current generative AI, how to educate the next generation about its capabilities, and how the global community should come together to address regulatory and societal issues. The risks of generative AI include the intentional creation of false information through large-scale language models to disrupt democracy and markets, according to Gary Marcus, an entrepreneur, former professor of psychology and neuroscience at New York University, and CEO of the newly established "Trusted AI Initiative." He also pointed out that generative AI's automated coding can alleviate the burden on overworked programmers. Wendell Wallach, Co-Director of the AI and Equality Project at the Carnegie Council for Ethics in International Affairs, highlighted the inequality exacerbated by generative AI between affluent countries in the northern hemisphere and impoverished countries in the southern hemisphere. For example, a blog published by the World Economic Forum in January 2023 noted that generative AI is primarily developed and used in the northern hemisphere. Generative AI extracts from training data in various languages. However, the most widely spoken languages naturally produce the most data. Therefore, those who generate AI applications using a large amount of data in a particular language are more likely to find useful applications, Marcus said. He stated, "Inequality is widening because those who use resource-rich, wealthy languages can do things that people of other languages cannot." Karishma Muthukumar, a cognitive science graduate from the University of California, Irvine, who focuses on using AI to improve healthcare, pointed out that many children learn about generative AI from peers or family rather than in school. She suggested offering a course on teaching AI applications. Muthukumar said, "This will require intergenerational dialogue and bringing together the greatest minds to find a truly effective curriculum." Many experts at the summit discussed the importance of community, ensuring that all stakeholders have a voice in discussions about generative AI. This includes "scientists, social scientists, ethicists, civil society, governments, and businesses," Marcus said. The International Telecommunication Union (ITU), established in 2015, is a United Nations agency responsible for developing and promoting global telecommunications standards and advancing telecommunications technology. Muthukumar stated that global platforms like ITU and similar conferences can foster a sense of closer connection, helping humans establish better relationships with AI. Wallach noted, "I hope that part of what this conference over the past few years has accomplished is to make people aware that this issue is on the table and to convey that awareness to our leaders so that they begin to recognize it as one of the problems we should not ignore." Regarding the ethical issues of using generative AI to address global problems, Muthukumar raised the question of what is considered good and how it is defined. She mentioned that the United Nations' Sustainable Development Goals provide a great framework and starting point to identify these goals and what can be achieved. Wallach pointed out that investing a significant amount of funds into generative AI companies does not necessarily solve the problems that AI for Good Summit aims to address. "One problem with the inherent value structure of the digital economy is that there are typically winners in each field," he said. "And the capital gains flow to those of us who own stocks in those winners. This poses significant challenges in terms of resource allocation for achieving sustainable development goals." He proposed that companies developing generative AI and other technological solutions to address global problems should also "take some responsibility to improve the shortcomings and trade-offs of the solutions they are choosing." The United Nations was also discussed. While Wallach commended the organization's efforts to bring stakeholders together to discuss global issues, he noted that the UN has a mixed reputation and cannot resolve "disharmony between nations." However, he believed that bringing the dialogue about generative AI and ethics to a broader audience would be beneficial. Ethical issues in AI can also vary depending on the context. "For example, the concept of fairness in AI varies greatly depending on its application," said Haniyeh Mahmoudian, Global AI Ethics Lead at DataRobot, an AI and machine learning software company, and member of the National AI Advisory Council, in an email interview with TechRepublic. "When applied to recruitment systems, fairness may mean equal representation, while in facial recognition environments, fairness may refer to consistent accuracy." Marcus believed that government regulation is an essential part of ensuring the future of generative AI. He said, "There is currently a tension between what's called promoting innovation and regulation. I think that's a false tension. We can actually promote innovation through regulation, telling Silicon Valley that you need to make your AI trustworthy and reliable." He compared the generative AI hype to the social media hype, where companies' growth outpaced surrounding regulations. "If we play our cards right, we will seize this moment - in individual countries like the United States and at a global level - where people realize that something needs to be done. If we don't, we'll have a year of regret," Marcus said. Source: https://www.techrepublic.com/article/un-ai-for-good-summit/

RECENT AI NEWS

RECENT AI TOOLS