A slight change of pace, or at least of topic, this week in Techzine Talks on Tour. We sit down with Reggie Townsend, VP of Data Ethics at SAS to talk about the important but not always easy topic of ethics and AI.

Townsend offers his and SAS's perspectives on what it means to build AI that truly serves humanity. "My job is to make sure that wherever our software shows up, we help people to thrive," in his own words.
Bias

The conversation tackles several topics around the ethical component of AI and data in general. One of them is the concept of bias. There is no such thing as an unbiased data set. As Townsend puts it: "Bias exists because humans exist." The real question isn't about eliminating bias entirely but distinguishing between harmful bias and purposeful representation.
Synthetic data

Another topic we discuss with Townsend is the use of synthetic data. This can be seen as both promising and potentially problematic. While it can fill crucial gaps in fields like clinical trials where finding enough participants with specific conditions is challenging, Townsend cautions against overreliance, noting the risk of creating self-referential systems that generate new biases.
Trust

A discussion about ethics, AI and data almost always touches the concept of trust too. We also discuss this at some length with Townsend. He specifically mentions the trust gap that is very real when it comes to AI. Addressing this requires not just transparency through tools like model cards (which SAS compares to nutrition labels for AI) but improved global literacy about these technologies.
Organizations should start bold

Looking forward, Townsend challenges companies to start bold when it comes to AI, data and ethics. We should look to address fundamental challenges. Townsend sees various opportunities in healthcare, mental health, but also around the topic of inequality.
Listen to this latest episode of Techzine Talks on Tour now.