Salesforce: Architects of Ethical AI Practices

In October 2016, Salesforce founder and CEO Marc Benioff informed internal employees, customers, and investors that Salesforce would be an AI-driven company. Earlier that year, Microsoft had released its Tay (for "Thinking about you") research chatbot project through a Twitter account. Microsoft shut down Tay after only sixteen hours because it had started to mimic the deliberately offensive behaviors of other Twitter users, and Microsoft had not trained the bot to avoid them. As Salesforce was interested in chatbots, Kathy Baxter, in her role at that time as principal user researcher, started looking into how the incident with Tay happened, and how AI-enabled system misbehavior could be avoided. In October 2017, when her job role expanded to user research architect, she was already working "on the side" within the company as an AI ethicist, educating fellow employees about ethical issues associated with AI and guiding product engineering teams in how to do their design work in ethical ways.