By Hugh Cadman, Interim Head of Content
The eruption of ChatGPT has triggered a major debate about generative AI and how it will reshape work and business, especially in PR.
A survey by a recruitment firm this year found 45 per cent of employees believe AI will help them do their jobs better. In the US, a Harris poll found 60 per cent of workers optimistic about the technology.
Although many of these respondents see the technology’s advantages in admin settings, ChatGPT’s runaway popularity is built substantially on its ability to create coherent content from scratch – and at speed.
Its success in attracting hundreds of millions of users has accelerated plans among the tech giants to come up with rival AI applications such as Google Bard and Palm, while Amazon and IBM have their own initiatives. More services are constantly evolving, such as Cohere Generate, or Copy.ai. Cohere, for example, will generate custom content for emails, advertisements, landing pages and so on.
OpenAI – the company behind ChatGPT – is in partnership with Microsoft, whose founder, Bill Gates believes AI “is as fundamental as the creation of the microprocessor, the personal computer, the internet and the mobile phone”. For Gates, the gains will be in productivity and more specifically in areas such as the development of new vaccines and medicines.
What are we left with in tech PR?
Given all the claims made for generative AI, it’s worth asking what its effect will be on B2B tech PR? At Whiteoaks our clients pay us for many services, including the creation of highly-targeted content written with skill to engage knowledgeable tech audiences.
But if AI automates marketing and creates comprehensible content in a few seconds, what are clients paying for? Should we get rid of our team of content-writers and ditch their years of expertise in favour of outsourcing to a ChatGPT-type service? Of course not.
Experience already shows that AI-created content, although coherent, needs human intervention to make it more tightly relevant to its purpose. It needs a writer to render it idiomatic so does not sound like every other piece of content put together by AI. It’s worth remembering that generative AI mines content from the internet. It will re-form what others have already said, which is hardly a formula for compelling originality.
And although it may have mechanisms to judge the authority of what it finds, AI is still far from infallible. We know that in the absence of figures or quotations, it can twist them or make them up. A business relying totally on AI-generated content may find that is repeating mistakes or biases that have gone uncorrected or unchallenged in other content already out there. Auditability and provenance are significant challenges, along with compliance with data regulation.
Commercial confidentiality and security concerns
There are serious concerns too about data leakage from generative AI services and the potential for breaches of commercial confidence through cyber attacks or lax procedures. It is not advisable to upload a mass of sensitive data of any kind into an AI service. Copyright is also a potential minefield, given that the AI is lifting and adapting pre-existing content and then claiming to create something new. Who does own the copyright? AI companies want to use the information inputted as part of their training models to hone accuracy – but for how long?
It’s not as if AI has come from nowhere. Apple Siri, Amazon Alexa and IBM Watson are all forms of AI and at Whiteoaks we have been writing about the technology for a long time – as a tool to transform the work of contact centres, for example.
Where does this leave us? We know the AI algorithms will continue to “learn” and improve the quality of generative AI output. But there will always be a gap between what a machine produces and the flair, insight and creativity of a human. A human with experience and client/market knowledge will be necessary to comb through copy to spot errors and the significant risks of legal, copyright and compliance infringements.
Where we should use AI, is to take care of the routine, repetitive element in content-drafting, which includes research and structuring. ChatGPT is a very useful research tool providing nuanced answers. And with skilled supervision, the technology is effective at reworking and refreshing existing content.
The advent of ChatGPT and advanced AI will prove to be watershed moment for many industries and will transform productivity in innumerable areas of work. It will create new jobs and lead to dramatically effective integrations we cannot currently foresee. But it will not eliminate the need for skilled and experienced content creation in B2B tech PR.
If this article has struck a chord – why not get in touch to chat?