AI ethics and the future of data

Since generative AI exploded into public consciousness, people and companies have been worrying about its ethics. But, says Amit Joshi, before considering rogue programs or the replacement of human thought, one must first consider the ABCs of AI ethics.

Accuracy is the first big issue. How can one trust or disseminate the work of generative AI when it is so prone to hallucinating?

Bias is a problem too. In a recent survey of image-generating AI models, when asked to produce a picture of a CEO, there was a 94% chance of a male image being produced. Requests for a “cashier” returned images of females 84% of the time.

The big C here is copyright. Generative AI may seem to produce original content but how is one to know where the ideas truly originate? Might that magical marketing slogan produced by Chat GPT simply be lifted from another company?

These problems are just the start. Mike Wade worries about the privacy of data, image recognition, and protection. Together, he and Amit discuss how big tech can mitigate these problems and whether a fair system can be achieved through regulation.

Find out more about IMD at imd.org