SALT LAKE CITY — With the gamut of unknowns swirling around generative AI — and its impact on communications — Method Communications is taking among the industry’s most dramatic steps to fend off potential problems by banning the use of the technology as a content creation tool until we reach a guaranteed safer place.

Method’s newly created policy “restricts the use of AI for research and inspiration in order to ensure that our work and communications and that we protect client confidentiality,” said executive VP Matt Stewart who, along with Tim Race, senior VP of narrative and thought leadership, is leading the initiative.

That goes for clients, too. “We are in the process of informing clients and updating our contracts to reflect our AI policy and ensure that all client content we receive which we in turn present to media is original as well,” Stewart (pictured) said.

The policy dates back to last spring when a client asked Method to submit original commentary to a reporter — which the client admitted was partially created by generative AI. Method deleted the elements created by the machinery before passing along the piece to the journalist, but was surprised by the author's admission nonetheless.

“At first, we thought he was kidding. This can’t be true,” Stewart said. “It was true.”

That raised a number of red flags for Method, not the least of which is that using AI-generated content could open up the firm to a host of problems, the most notable one being plagiarism.

Though ChatGPT doesn’t outright copy other people’s work, it’s not risk-free, either. As Race — a former New York Times technology editor who has built out Method’s editorial team — puts it: “We have to recognize that AI is derivative. Generative AI is just scanning what has already been thought and said.”

Plus, the content ChatGPT produces isn’t up to Method’s standards.

“We are facing brilliant reporters and they need to serve really smart readers who expect A-plus content,” Stewart said. “It’s mostly churning out C work,” he said, adding that while AI-generated content will likely improve with time it currently is “not anywhere near a final product."

“It’s just not good enough,” he said.

Other concerns include the risk of breaking client confidentiality by “feeding the beast of AI” information that winds up in someone else’s ChatGPT, Stewart said. Using ChatGPT for content would also preclude junior staff from garnering the skills they need to advance.

“People need to learn how to write amazingly. They need to think and research. AI can be part of that, but I don’t want to outsource creativity or stringing a brilliant pitch together because then you’ll never learn,” he said.

Method’s policy is stricter than what the PR Council recommended in its guidelines issued in April, which do, though, urge PR firms to take a cautious approach to using the technology by fact-checking content; taking steps to avoid copyright infringement and plagiarism; being transparent in use of generative AI; and having training and best practices in place.

The PRC guidelines also pay particular attention to the danger of bias in AI-generated copy and imagery, and the importance of the PR industry using best practices to make sure biases in material are neither overlooked nor distributed.

For Race, generative AI is missing the one component that is the basis for the thought leadership that makes up a good portion of his group’s work: humanity. “There is just no way that it can come up with original executive thoughts,” Race said.

“The best business thought leadership comes from the experience and expertise of executives in the real world. Generative AI has never lived a day outside its silicon synapses.

“Until you have bots that are observing and acting in the real world, I don’t think you are going to get true and strategic insights,” he said.

None of which is to say Method is shunning generative AI. It's just, at this point, the agency is reserving the use of it for the likes of  research and idea generation. Stewart said AI is also proving to be an “immensely powerful” application for PR reporting, with the promise of being able to handle the multitude of tasks currently handled by different tools.

And Method’s policy isn't the firm's be-all end-all statement on generative AI; Rather, the policy is continually under review so that it changes and takes shape in keeping with the technology.

“It’s not like we’re against technology. We are a top tech PR firm, so we love new stuff that is going to make the world better and innovation,” said Stewart. “We are just not diving into the pool blindfolded.”