Emma Watson has built up a fierce reputation as a force for good in the world. She’s an active campaigner for equality, has a number of charitable engagements and after all, did help defeat evil over eight Harry Potter movies. Few celebrity reputations come higher.
Imagine my surprise then, when listening to what appeared to be a recording of Watson proudly reciting sections of Mein Kampf. Except, she wasn’t. She was the victim of a terrifyingly realistic AI deepfake.
Watson isn’t the only high-profile figure to have been hit by this increasingly common threat either. There was Donald Trump being brutally arrested (this time, not for real) and who can forget striking images of the puffer-jacket-Pope? As amusing as these might be, they represent arguably the greatest threat to personal and business reputation since the inception of social media.
Crucially, the advancement of generative AI technology has lowered the barrier to entry for malicious actors. Whilst to a certain extent creating fake images has always been possible with photoshop, this required a great deal of technical expertise, skill and practice. With AI, you need none of those attributes. A computer and an imagination are the only essentials.
For example, Microsoft claims it needs just a three-second sample to recreate someone’s voice with their AI, VALL-E. Suddenly, anyone with any sort of online presence is at risk.
And sadly, the world has never been short of people with nefarious aims. All the way from state actors to trolls in their mother’s basement. The difference is, now they have access to easy-to-use technology that creates content that could fool even a canny eye. Newspaper picture desks are apparently spending hours sifting through images they receive to check for AI generation, and it’s only going to get worse.
At present, responding to the AI deepfakes is best ignored or handled with a smile and a joke; clarifying that they aren’t in fact dating a llama or that image of them dining with Tupac probably isn’t real. However, if these continue to progress in likeness and authenticity then a more severe PR response will become increasingly necessary, with speed in clarifying misinformation of the upmost importance.
When you ask AI to mark its own homework, querying what effect it will have on Reputation management, it gives you a predictably chipper answer:
“AI brings significant benefits to reputation management. It empowers PR professionals with advanced sentiment analysis, real-time media monitoring, proactive crisis management, and personalized communication. By harnessing AI, PR teams can enhance their ability to monitor, analyze, and respond to public sentiment, protecting brand reputation and building stronger stakeholder relationships.”
And to be fair to it, these are good points. Monitoring media sentiment and picking up on coverage are effective tools underpinning any communications strategy. Advancements in these areas, as well as predictive tools, will be key weapons in our armoury to help protect our clients. But right now? I’d sooner ask a cat to bark than ask ChatGPT to manage ‘proactive crisis management and personalised communication.’
In addition, what it cannot replace is the uniquely human ability of providing strategic counsel; understanding people’s motivations and perspectives; and offering sound judgement on how a situation may unfold. Calm heads in a storm are difficult to replicate.
All of this also highlights the importance of proactive reputation building. To go back to Emma Watson, whilst I found it shocking, part of the reason I knew it couldn’t be true is because of the years of reputation credit she has built in the bank.
So, if audio is released of you reciting Mein Kampf that you don’t remember doing yourself, you might want to consider ringing a PR professional, and not punching your problem into ChatGPT.