What ChatGPT Reveals About the Urgent Need for Responsible AI

Concerns about power, governance, unintended impacts on customers, capability overhang and ill-defined copyright issues plague generative AI.

First there was a mad scramble to incorporate AI into the processes and operations of organizations; now, organizations are recognizing the problems that can emerge with the use of AI, particularly generative AI. Consider ChatGPT: The program can generate credible-looking, but often incorrect, scientific texts, it can parrot hate speech, and can proliferate biases in avatars and images. But these aren’t the only issues. The data sets used by generative AI contain a wide variety of materials, laying the foundation for copyright claims. Though smaller organizations can use AI, only very large corporations with access to resources like large data sets, computing and engineering power can build generative AI. This means the power of generative AI is centralized among large, powerful organizations, leaving marginalized...