The way we communicate with each other has come a long way, from the printed word to online. In that time, audience attention span and media consumption shifts have been made possible by new technologies. These changes have diversified our industry’s work and brought new challenges and opportunities for both PR professionals and the journalists that we work with.
The latest of these transformative technologies has been somewhat of a hot topic over the last few weeks: ChatGPT, a chatbot capable of writing text in response to human prompts.
AI and automation have long been used by businesses looking to reduce administrative hours, combat staff shortages and free up employees for more creative or strategic tasks. The recent launch of the ChatGPT prototype and the Lensa AI portrait app has brought a shift in the expectation of what AI can deliver. In fact AI is making its mark as a truly disruptive technology: just last week, Microsoft announced its plans to integrate the AI into its Bing search engine.
As the latest challenge to the established norms of the communications industry, how should we approach AI-copywriting?
What is ChatGPT?
For those who don’t know, ChatGPT stands for Chat Generative Pre-trained Transformer, and is a chatbot capable of writing “impressively detailed” and “human-like” text across a whole range of formats. Whether a user wants business advice from the point of view of Elon Musk or a Shakespearean sonnet, this AI is intelligent enough to create totally new content by leveraging millions of texts on the internet.
In PR, it proves to be a helpful tool. It can generate new social media prompts, make research far simpler, and it makes easy work of localisation and translation tasks with 97% language processing accuracy. As we often say of AI in other industries, these capabilities could free up PR professionals to focus their time on more creative and fulfilling tasks – but with Generative AI now capable of creating novel content from a simple prompt, what does it mean for the industry?
Generative AI: the ultimate thought leader?
Fast learning with the ability to adapt from every spokesperson before it, ChatGPT has the potential to be the ultimate thought leader on any topic – or so some people think. But the reality is that there is a difference between writing convincing copy and creating meaningful, sensitive and valuable content that inspires trust and – perhaps most importantly – avoids legal and reputational issues. This is because AI content creation, as it stands, still carries a few important drawbacks:
Firstly, an AI that learns from existing content and ideas may open brands up to accusations of plagiarism and potentially copyright infringement – the lack of a paper trail on ChatGPT’s sources should truly be a concern for brands.
Secondly, businesses often need to communicate with a level of nuance that only humans can comprehend, and that PR professionals are trained to recognise. The importance of sensitivity is especially clear over the last few years, as brands communicated through a pandemic, a cost-of-living crisis, a recessionary climate, and in times of geopolitical strife. It is very difficult to train an AI to be wary of the millions of contextual points that humans are innately sensitive to, and a simple mistake can be very damaging. Just last year, KFC had to apologise after an AI designed to generate push notifications around national holidays mistakenly suggested that customers celebrate the anniversary of Kristallnacht.
ChatGPT, which learns from an expansive database of online texts, also carries the potential to perpetuate the very worst facets of the internet.
It is well documented (and criticised) that sensationalised news stories generate the most engagement online, and that social media algorithms tend to favour divisive or extreme content. If a chatbot is drafting a social media post about a nation-state ransomware attack on a hospital, will it prioritise popularity as defined by these algorithms, or true editorial value? It is also foolish to ignore the ways in which many of these algorithms have been shown to have actively radicalising effects on their users. As Microsoft’s 2016 chatbot demonstrated quite clearly, AI’s ability to learn quickly can create real problems if left unchecked.
These limitations may be addressed in the future, but it is difficult to imagine teaching nuance to an AI without also instilling the implicit biases of its designer. Until then, the concerns about whether ChatGPT delivers the right information have led to the software being banned in the writing of academic papers, by whole school districts, and even by forums such as Stack Overflow. PR Professionals must toe the line between embracing new technologies and staying wary of our due diligence as our clients’ representatives.
And now I’m afraid I have to turn to an industry cliché to lay out the last barrier to entry for AI thought leaders: people buy people. While AI journalists and content creators are now a possibility, what is still unclear is how much trust the public will have in their opinions. Brand spokespeople have demonstrable and current experience within their industries and can speak to community challenges with the whole context of their expertise. They can speak at events, shake customers’ hands at expos, and will be far more compelling advisors for the many professionals looking to them for wisdom.
The many uses of ChatGPT are still up for discussion, and I am sure that it will transform multiple industries and prove a useful tool to those who take the time to learn how best to use it. As these technologies learn and improve their output even more, it is important that we in the creative industries keep informed and open-minded on the role that they might play in our work. But for now, when creating brand content and messaging strategies, there is still more than enough room for us humans as holistic thinkers, sounding boards, and trusted consultants.
For more information about how C8 Consulting can help to create compelling campaigns and kick-start conversations for your business, please feel free to contact us.