In the world of increasing crisis threats and preparation, a new one has come to the fore, this one driven by AI; but perhaps not one that’s on the communicator’s radar yet.
A recent article in the May 11 Axios AI+ newsletter noted how in the trial brought against Open AI by Elon Musk, AI internal chats with chatbots can be viewed as a diary of sorts and could become evidence in a court of law. As the article states “Courts are increasingly treating chatbot conversations as discoverable evidence, raising new legal and privacy concerns for AI users.” And cases from earlier this year have served as legal testing grounds on how AI conversations can be used in matters of litigation.
Even though Musk’s claim against Open AI was rejected by a jury because of the statute of limitations, it illustrated the new crisis threat.
The Evolution of Chat and Legality
It used to be social media outrage and leaked or subpoenaed emails, texts and meeting notes could point to potential internal issues and decisions. Now senior leaders’ own conversations with an LLM could be used to argue guilt or innocence. Now not only have lawyers been put on alert, but so should communication professionals.
Obviously, demonstrated intent to break the law can’t be “spun” no matter how good your PR team is. And it’s common knowledge among PR pros that AI-derived advice isn’t always the best solution. At the same time, it’s critical to remember AI chats don’t just disappear once they’re ended. They’re logged in a database that’s then used for LLM training. And if they’re logged and saved, they’re recoverable.
“The new chatbot or AI diary threat is not new at all,” says Heather LaMarre, CEO of The LaMarre Group, a strategic communications firm. “The age-old adage, ‘be careful what you put in writing’ still applies. Discipline matters in executive communication.”
Steps Communicators Should Take Regarding AI Chat
So what do communicators need to do? First, do your own research. Has there been any past internal behaviors or queries involving chatbots that could become litigious? Share your findings with legal so they don’t get blindsided and you’ve become allies on the issue.
The next step should be obvious. Rule #1: be careful what you put into a chatbot.
The third step should be to educate senior leadership about the issue (see Rule #1). The same holds true for practitioners. It’s not just about the prompt you’re currently writing, it’s about the intent.
For example, are PR professionals asking a chatbot to come up with a rationale for a strategy? Potentially, that could raise questions about fairness, ethics, stakeholder privacy and other issues that could wind up in court. This is not to say practice self-censorship, but it does add another layer to consider.
This is where PR and legal should continue to collaborate on not just what the legal issues are, but also how they overlap with the strategic responsibilities that call on PR’s expertise and counsel: reputation, credibility, competence, character, internal culture, ethics, etc.
Finally, use strategic communications risk management tools to strategically plan for this scenario, just as planning for other crises would be done, whether they’re natural disasters or man-made/self-inflicted crises.
“As AI chats become commonplace, risks around privacy will continue to grow and basic communication risk management practices will be essential,” LaMarre notes. “Unforced errors such as these are another reminder of why CEOs need executive level communications advisors.”
In today’s business and social environment, PR pros can’t inoculate their organizations against this growing legal trend and crisis threat, but they can prepare for it.
Gregg Feistman is Professor of Practice, Public Relations at Temple University