‘A starting point for creativity:’ How PR pros are testing ChatGPT

They’re also heeding warnings that the tool could create ‘a cavalcade of bullshit.’

Young staffers are spearheading agencies' forays into ChatGPT. (Photo credit: Getty Images).

When Facebook launched in February 2004, it took the platform roughly 10 months to amass 1 million users. Nearly two decades later, OpenAI’s ChatGPT achieved the same feat in just five days

Admittedly, the world is far more digitized than when Mark Zuckerberg created the social media giant. Nonetheless, the buzz around ChatGPT has been infectious, forcing almost every industry to pay attention

Read what happened when PRWeek asked ChatGPT to address users' concerns about AI chatbots.

Its impact on the workforce is undeniable, as well: a ChatGPT application was recently shortlisted for an interview by communications consultancy Schwa’s recruitment department. 

In an industry so keenly concentrated on emerging platforms, PR executives are dipping their toes into ChatGPT’s uncertain, yet exciting waters. But where do they even begin? 

San Francisco-based Mission North is dabbling with how ChatGPT can jumpstart the creative and content-development processes, according to co-CEO Bill Bourdon. 

“We’re in the research and experimentation phase right now,” Bourdon says, adding that Mission North has pulled together a “working group” on AI technology, derived predominantly from its content team. 

Mission North’s group is exploring how ChatGPT can serve as an organizational tool, pulling together background notes or statistics on publicly available and previously documented information. 

“We view a potential future that has great writers and creatives using a platform like ChatGPT to spend less time on the busywork of commodity content creation,” Bourdon says. 

A recent wave of AI-powered tools, including Dall-e and Jasper, had previously captured the attention of junior to mid-level agency staffers, says Rob Stone, head of Z3, Zeno Group’s global Web3 consultancy.

“It wasn’t something where a conversation started at a senior level and then we started asking people to look [into these platforms], it actually happened the other way around,” he says. 

The numbers back up Stone’s experience. According to data from networking app Fishbowl, 27% of professionals have used ChatGPT to assist with work-related tasks, with the highest rate of adoption among Gen Z staffers at 29%. 

Armed with experience from similar AI models, Z3 is examining ChatGPT’s ability in copywriting and creativity. 

“We will quickly start to explore what [ChatGPT] means for our clients and, strategically, where we need to go,” Stone adds. 

While ChatGPT’s capacity is staggering in terms of helping to eliminate time-consuming and mundane tasks, it cannot yet replace the critical human element, argues Axicom global president of digital Brian Snyder. 

Snyder compares generative AI to other tools Axicom uses, such as Brandwatch for social listening or Onalytica for b-to-b influencer marketing, which provide “raw materials” and opportunities for staffers to capitalize on through social media or media pitches. 

While a well-constructed ChatGPT query could build a foundation for a blog post or press release, it doesn’t yet understand context such as specific audiences, comms and business objectives or brand voices, Snyder adds. 

“Our content creators use ChatGPT to provide a starting point for creativity. It gives you the clay you can then mold,” he says. 

Axicom recently supported the creation of an AI-powered, editorialized and journalist-reviewed report on CES 2023, entitled Qant CES Report 2023, as well as daily newsletters throughout the conference, according to a company statement. 

The report, built using the model behind the ChatGPT interface, GPT3, is a part of Qant: l'avenir du numérique, a French professional newsletter entirely written and illustrated by robots, under the control of two journalists.

“Generative AI from Qant provided the raw materials for the coverage, while the journalists worked with the tech experts in our Paris office to review, contextualize, verify and then turn it into these newsletters that we put out everyday during CES,” Snyder says. 

The full CES report will be available on Axicom’s website on January 31, with an accompanying presentation in Paris around the same time. 

Mike Houlihan, EVP at WE, echoes Snyder’s sentiment on human involvement. 

“I talked to one researcher who said, ‘With ChatGPT, the outputs can all sound like pop music: it all sounds the same,’” he says. “The question is how do we make it sound like Mozart? That’s where communicators add value. We can add meaning, nuance and differentiate the content.”

Moreover, ChatGPT’s responses, and ultimately effectiveness, is determined largely by the prompt it is fed, Stone notes.  

Even Sam Altman, CEO of OpenAI, warns against an early overreliance on ChatGPT. In early December, Altman tweeted, “ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. It's a mistake to be relying on it for anything important right now.”

Few disagree that ChatGPT has the potential to become a powerful tool in any comms practitioner’s toolbelt, but PR leaders do have reservations about misinformation. 

“From a PR perspective, there’s potential for a cavalcade of bullshit,” Snyder says. “Think about misinformation at scale, astroturfing at scale or being able to generate 1,000 letters to an editor or congressperson.”

Read what happened when PRWeek asked ChatGPT to address users' concerns about AI chatbots.

Problems regarding accuracy, bias and legality could also arise, Bourdon notes. New York City’s public schools recently banned students and teachers from using ChatGPT over concerns about cheating on assignments and the spread of misinformation. 

“The same issues [for the education system] stand for communications and media companies. How do we develop competent writers and communicators with a tool like this and avoid cheating?” Bourdon asks. He adds that ChatGPT will need to build in its own copyright protections to scale. 

As ChatGPT rapidly develops, with a projected $1 billion in revenue by 2024, Snyder poses a few questions to his fellow communicators. “Will we as PR practitioners be trying to influence people as we have since the beginning of this profession? Or are we going to be trying to influence the AI that people rely on?”


Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Register
Already registered?
Sign in