Media training tip:
- Don’t say yes to an interview if you won’t answer obvious questions
- Answer part of the question
- Never say “pass” or “no comment”
The WSJ’s four-part podcast on OpenAI, titled Artificial is great. Yesterday the last episode dropped and it had an interview with Helen Toner, a former OpenAI Board member who had a hand in pushing out CEO Sam Altman. Toner presumably gave an interview to the WSJ to defend her reputation, but her performance on the podcast is head-scratching. She won’t answer questions.
For a reporter to get an interview with Dr. Toner is a huge get. Rigorous questions will be asked. Yet, Toner declined to answer even some routine questions (minute 14):
Kate Linebaugh (WSJ journalist): What was Sam Altman’s relationship with the board?
Helen Toner: Sam … yeah, I think probably better not to get into this.
Kate Linebaugh: There were a lot of things Helen didn’t want to get into in our interview. Was there concern that OpenAI was on the precipice of AGI?
Helen Toner: Pass.
Kate Linebaugh: Did you tell the board about this paper or think you needed to?
Helen Toner: Yeah, I think I’ll just pass on saying too much more about the paper.
Toner is affiliated with Georgetown and wants to ensure AI, and certainly AGI, is safe for humanity. It’s hard not to be on her side. But why give an interview like this if you won’t answer likely questions? You never need to say “pass” or “no comment” to a reporter. It makes you look unprepared, even suspicious …as though you have something to hide.
Perhaps Toner felt obligated to answer each question, in its entirety, as they stood. But she certainly is under no obligation to speak for the entire Board. She could have taken the first question – What was Sam Altman’s relationship with the board? – and refocused it, narrowing the parameters until she stood on solid ground. She could have said:
“I can’t speak to Altman’s relationship with the Board. I can speak to my relationship with him…(and then portray what was clearly a complex relationship).”
Then the second question: “Was there concern OpenAI was on the precipice of AGI?”
Toner certainly doesn’t have to go there, nor should she. OpenAI’s proximity to AGI is certainly a trade secret. But she could have taken the higher ground and said: “In these situations it’s rarely one thing that causes a Board to rethink the CEO. It’s often a culmination of incidents and experiences.” Would that not have been better?