Issue: Q4 2022
Join CIPR
CPD Points
The back of person’s head as they are looking at three computer screens.
Image created using Midjourney
TECHNOLOGY
6 minute read

Faking it

Move over the metaverse, say cheerio to crypto and nip those NFTs in the bud: the latest buzz in the tech right now is generative AI. From bespoke statements to magically conjuring up press releases, Christian Koch reveals how it could be transforming the world of PR very soon…

While the debate still rages around their effectiveness, the press release is still a core part of every PR's arsenal of skills: extracting facts and info from a client, before crafting it into a newsworthy message that will pique journalists' interest the second it drops in their inboxes (or maybe not). 

Very soon though, this task could be taken out of PRs' hands. In the last year, online ‘press release generators' and text generation engines such as Jasper, GPT-3 and copy.ai have soared in popularity (PRs are using them for blogs and emails too). If you'd previously assumed these artificial intelligence (AI)-powered applications are incapable of, say, capturing the nuances of a killer ‘top line' or stripping away superfluous waffle, think again: Jasper (recently valued at US$1.5bn/£1.25bn) can compose convincing press releases with just a few simple text prompts. 

It's all part of the still-erupting trend for generative AI, currently whipping Silicon Valley up into a frenzy of excitement. Broadly speaking, generative AI is artificial intelligence that creates original text, images, audio, videos and code without using pre-existing data sets (ie video footage). In recent months, a spate of new generative AI services has reconfigured the digi-sphere, meaning this technology is in the hands of consumers. At this very moment, people are creating videos without using cameras, audio without using recordings, and yes, content such as press releases without much human input at all. Just a few hastily written prompts and voila! - your press release is ready. 

"We're on the cusp of an AI-led revolution and paradigm shift that will transform the future of content creation" Nina Schick

"AI-generated content is going to become universal; pretty soon all of the digital content in our ecosystem will be generated by AI," says Nina Schick, director Tamang Ventures and author of Deepfakes: The Coming Infocalpyse who recently delivered a well-received talk on the subject at the CIPR Annual Conference. "We're on the cusp of an AI-led revolution and paradigm shift that will transform the future of content creation, and the way we think and feel," she told conference-goers. 

A quick scroll through TikTok or other social media feeds will reveal how ubiquitous generative AI has become recently. You may have noticed some surreal memes - Kermit the Frog in the style of Edvard Munch or Shrek as Mona Lisa, say. Some are so persuasive, Schick predicts "teenage YouTuber/TikToker will sit in their basements creating content just as sophisticated as a Hollywood studio. Within the next two-three years, you could make feature-length films on your laptop."

 

FAST AND FURIOUS 

The rise of generative AI has been startlingly rapid. Last year OpenAI, an Elon Musk-backed research lab, launched neural network Dall-E, which was trained on 400 million images and can create a unique image from a simple text phrase such as: ‘Tom Cruise eating tulips'. Other players such as Midjourney soon followed. In October, the UK-based Stable Diffusion reached 10 million users. Not bad for a company launched just two months before. Indeed, Schick estimates generative AI has already created three unicorns (startups with a $1bn+ value) in the last year alone.

Their success has triggered a "generative AI arms-race" according to Schick. Microsoft is integrating Dall-E into its Designer app (you could see text-generating applications pop up in Microsoft Word), Meta has unveiled its Make-A-Video service (which generates video from text) with Google's Imagen Video system to follow. Indeed, Schick believes 90% of all video content will be synthetic (ie ‘fake') by 2025.

What does this mean for the PR industry? Well, aside from magically manifesting press releases in a few seconds-flat, perhaps the biggest boon is it could remove monotonous admin for PRs. Instead of wading through turgid trend reports, generative AI could extract key facts for PR teams within seconds. It could also create targeted media lists, help brainstorm creative ideas, spruce up press releases with snazzy infographics and videos, create content for social media accounts, identify influencers in a specific field, craft pitches, generate reports on media coverage and analytics, plus create personalised content for each journalist. 

 

THE DARK SIDE 

Even if PRs choose not to use it, generative AI could intersect with their business in other ways, most notably as a security risk. The rise of remote-working and Zoom job interviews during the pandemic saw the FBI recently issue a warning that criminals were using deepfake job candidates to gain access to corporate information. 

Deepfaked videos of clients (such as a celebrity or CEO) could create crisis comms headaches for PRs by duping legions of social media users. Part of the problem is that much generative AI can appear alarmingly accurate. It creates an "existential, societal risk" says Schick. "If we're looking at a digital ecosystem where everything can be faked, including real people cloned by AI, how do we know anything is authentic or real [including] the media?"

Just imagine if a faked video of a CEO resigning appeared, which could send markets into a panic. Some firms have already fallen prey: in 2019, a UK-based energy firm paid nearly £200,000 into a Hungarian bank account after being phoned by a scammer using an apparently deepfaked voice impersonating the CEO's voice.

The potential for disinformation isn't the only ethical issue. Generative AI also raises problems of copyright issues (what will AI-artwork mean for artists/designers?) and racial/gender bias (because the AI is trained on images pulled from the internet, Dall-E 2 noting if you typed in the word "nurse", pictures of a woman would invariably flash up; the same thing happened with "CEO" and images of white men).

While tech and users grapple with these concerns, Schick's advice for PRs and other professionals is to embrace generative AI, rather than reject it. "Generative AI is a tipping point," she says. "Rather than something that will be integrated into society, it'll change the framework of society itself. We should absolutely seize the creative potential of this moment. Go and have a play on Jasper, but [equally] think about the implications of what it means for humanity too."

For more CIPR event information, go to https://www.cipr.co.uk/CIPR/Events/Event_Calendar.aspx