RightsCon, the world’s largest digital rights conference, recently took place in Costa Rica, bringing together internet ethicists, activists, and policy makers to discuss artificial intelligence (AI). The discussions about the risks of generative AI at the conference differed from the warnings coming from Silicon Valley. While tech luminaries have called for regulation and urgent action to address the “existential risks” of AI, speakers at RightsCon pointed out that the rush towards AI is driven by corporate profit-seeking rather than regulatory clumsiness or technological inevitability.
The rapid deployment of large language models without risk assessments, disclosure of training data and processes, and lack of attention to potential misuse is concerning. In a session at RightsCon, Gideon Lichfield and Urvashi Aneja criticized Google’s actions, suggesting that the AI gold rush is a result of competition between companies. Kent Walker from Google defended the advancements in AI, emphasizing the potential social benefits.
AI researcher Timnit Gebru questioned the notion of AI’s existential risks, arguing that attributing agency to AI tools is a mistake. She also expressed frustration with the narratives surrounding AI, pointing out the inconsistencies in how AI is portrayed by those who have invested in AI companies.
Experts like Frederike Kaltheuner highlighted the known harms of current AI use. Recommendation algorithms on social media platforms have been shown to prioritize extreme and misleading content, and large language models contribute to the spread of persuasive misinformation. Kaltheuner also raised concerns about the inappropriate use of generative AI chatbots in contexts like psychotherapy, as well as the environmental impact of the computing power required to run large language models.
Amidst the discussions about the risks of AI, Kaltheuner questioned whose extinction is being discussed when it comes to the future of humanity. She argued that marginalized communities are already experiencing harm, making the focus on global human extinction seem cynical.
It is clear that the risks associated with AI are not just speculative but are already causing real harm. Addressing these known risks is crucial, rather than solely focusing on potential future risks. The discussions at RightsCon emphasized the need for greater oversight, transparency, and responsibility in the development and deployment of AI technologies.
