Who thought this was a good idea?

Wrong Occasion

If you're going to send an official email on a sensitive and grim topic, you probably shouldn't generate it using AI. That's the lesson that two deans at Vanderbilt University learned the hard way last week when they thought (for some reason) it was a good idea to send a ChatGPT-generated email to students concerning the recent mass shooting at Michigan State University, the student newspaper Vanderbilt Hustler reported — which promptly landed them in hot water.

The unfortunate email was sent out on Thursday from the Peabody Office of Equity, Diversity, and Inclusion (EDI) at Vanderbilt's Peabody College, and is about as vague and platitudinous as you'd expect from text composed by an AI that isn't even up to date with current events of the past two years.

"The recent Michigan shootings are a tragic reminder of the importance of taking care of each other, particularly in the context of creating inclusive environments," the email reads.

After a bunch of empty prose, the email concludes with an almost comical footnote: "Paraphrase from OpenAI's ChatGPT AI language model, personal communication, February 15, 2023."

Need For Humanity

As Vanderbilt Hustler notes, the email does not use any Vanderbilt-specific terms other than "Peabody," and refers to multiple recent mass shootings when there had only been one. It's clear, too, there aren't any specific tie-ins to the grim event in question.

Credit to the student newspaper, though, whose reporting forced the university to respond.

"While we believe in the message of inclusivity expressed in the email, using ChatGPT to generate communications on behalf of our community in a time of sorrow and in response to a tragedy contradicts the values that characterize Peabody College," said Nicole Joseph, an associate dean at Peabody's EDI office responsible for the AI-email, in an apology email.

Peabody College's dean of education and human development Camilla P. Benbow said in a statement that her office was unaware of the email before it was sent, and that she was "deeply troubled that a communication from my administration so missed the crucial need for personal connection and empathy during a time of tragedy."

As a result of this collegiate contretemps, both Joseph and EDI assistant dean Hasina Mohyuddin, who was also responsible for the email, will "step back from their responsibilities," according to Benbow.

Needless to say that, as generative AIs become more ubiquitous and powerful, it's a race to the bottom to see who can come up with the dumbest and most ghoulish ways of using them.

More on ChatGPT: Bing AI Says It Yearns to Be Human, Begs Not to Be Shut Down

Share This Article