Close

Algorithms, data, and platforms: the diverse challenges of governing AI

Mark Nitzberg (right) and John Zysman (both University of California)
Mark Nitzberg (right) and John Zysman (left), both University of California

With the introduction of ChatGPT and similar tools based on artificial intelligence (AI), public debates have sparked renewed interest about the challenges and opportunities of artificial intelligence. In this article, Mark Nitzberg and John Zysman discuss how AI can and should be governed from a national and global perspective. They argue that challenges of governing AI cannot be considered in isolation but in the larger context of a ‘toolbox’ that includes algorithms, data, processing power. And even more importantly, both platform firms and the platform technology itself need to be governed because they generate big data on which AI tools can operate. To strike a balance between encouraging the potential while minimizing the risk of AI, according to Mark and John, governing this toolbox needs to focus on two complementary approaches. One needs to focus on sector-specific applications to account for variations in gains and costs depending on the purpose and domain where it is applied. The other approach should focus on overarching principles and rules that cut across many social domains and economic sectors. However, since AI is identified as a critical component for national success and a global debate on AI beyond ethical expressions is unlikely, the article concludes that rather than a mutual agreement on a set of goals and market/social rules, governing AI should have the objectives of interoperability amongst nations with sometimes fundamentally different political economies.