Everyone’s talking about it…ChatGPT is in the headlines, in the boardroom, on the editing floor and being discussed at all levels across every agency. Including Midnight.
At the time of writing this blog, the FT had just hired its first AI Editor. Google had just warned the world that ChatGPT could deliver ‘convincing but entirely fictitious answers’ and there were reports that $100bn was lost from Alphabet’s market value after a recent AI search event.
The big questions on everyone’s lips are: will AI copy replace jobs? What are its limitations, risks and opportunities? What are the best uses for it? Which sectors could benefit from it the most? What does the future hold?
We asked our most senior personnel for comment and talked to other agency owners to get a better understanding of the current state of play and the implications for AI copy in the wonderful world of Public Relations and marketing.
ChatGPT will become a powerful tool for business
Our Account Director Hydar Al-Bedaery, who’s been using ChatGPT since its inception had this take: “Chatbots have the potential to revolutionise the way businesses interact with their customers and will become a powerful tool for companies in the coming years.
“However, it’s important to note that chatbots such as ChatGPT are still in their early development stages and will only partially replace human thinking. Currently, they can handle simple, repetitive tasks, and the core output they generate is for good desk research rather than publishable pieces.
“It’s also worth noting that chatbots require multiple inputs to feed the correct output. The accuracy and suggestions may only sometimes be accurate to current affairs due to the limited knowledge of events that occurred after 2021 in the current database that ChatGPT has access.”
Could ChatGPT and AI copy replace jobs?
The simple answer to this is yes, it already is. FT had just hired its first AI Editor. Surely more will follow.
We’ve heard tales of tasks usually given to paralegals being conducted by AI such as summarising long court reports or other legal documents. We spoke to another agency owner recently who claims to save upwards of three hours per task by giving editing jobs to AI to cut down long documents to specified word counts.
Will it replace jobs in PR? Hydar commented: “Chatbots have established themselves as a fixture in the digital and PR landscape and will undoubtedly progress in sophistication and prominence in the years to come.
“Their capacity to minimise the cognitive load involved in producing written materials will be a key driver of their increased adoption. Moreover, the advent of chatbots may usher in a paradigm shift in our approaches to discovering and generating written content, but they won’t replace humans.”
What are its limitations?
When you login to ChatGPT it warns you of three limitations from the outset:
- May occasionally generate incorrect information
- May occasionally produce harmful instructions or biased content
- Limited knowledge of world and events after 2021
Despite this, we know some agencies are using it to generate copy for their clients’ social media posts and even to write press releases. Our team of trained journalists can produce copy quickly and efficiently without the need for a chatbot. However, using AI for desk research and editing is useful.
Producing copy via AI could also have negative implications for SEO. If your website blog content has been produced via AI and Google (or another search engine) can spot this, it might penalise you for it. However, Google’s position on AI is evolving (perhaps unsurprisingly since it launched its own AI chatbot) and essentially the advice is to make sure your content is written for people, not for search engines, and to avoid ‘spammy automatically-generated content’.
So… if you’ve got to input a lot of information to get the chatbot to spit out what you need, and there’s a risk of that information being factually incorrect, and you’ve got to re-write a lot of it, wouldn’t it have just been quicker and easier to write it yourself? At the moment…that’s where we’re at. But we’re keeping a watchful eye.
What are the best uses for ChatGPT?
At Midnight, we are looking at this as an effective editing and research tool. The copy output isn’t good enough yet, although having a starting point is often useful (so long as that information is fact checked and edited).
So, for example, if we had a client send us an incredibly long document that we needed a quick summary of, using ChatGPT could be very useful. Or, if we had produced an incredibly long document that needed a lower word count, we might use AI to assist.
If we’re researching ways to make a new campaign sing or conducting competitor reviews for clients, using AI for that research is useful, so long as you don’t need information post 2021.
Should AI copy be regulated?
Does AI copy such as ChatGPT need regulating? Potentially. Especially when it comes to copyright. In the world of PR, we have to be very careful not to breach the NLA and CLA licenses which protect the copyright of newspaper and magazine publishers. Will AI copywriters really create unique copy that doesn’t breach copyright? This remains to be seen.
Which sectors could benefit from AI copy?
Any sector which requires editing or desk research assistance or simple copy jobs could benefit from ChatGPT and the like. The FT won’t be the last publisher to appoint an AI editor. Law firms will surely use it for simple document summaries and other simple tasks. PR and marketing firms are already using it for a variety of tasks. Our local coffee shop in Hove uses ChatGPT to produce their menus and recruiters are apparently using it to write job ads and reply to candidates. Could ChatGPT re-write your CV? This journalist tried it with disastrous results. Not advised.
The bigger question therefore is probably what sector wouldn’t benefit from ChatGPT and other AI tools?
What does the future hold for AI copy?
Technology moves at a rapid pace. Just look at how quickly we’ve moved since Midnight was originally founded in 1995 – a time when press releases were sent out via fax machines, photo stories were delivered via bike courier to the papers and Google wasn’t even launched yet. Fast forward to today when we’re using market-leading technology to analyse our campaigns, appointing digital-skilled personnel and trialling AI tools.
We pride ourselves on always being ahead of the curve, and we’re cautiously excited about AI and the potential for our sector and others. However, with all new technologies the best approach is to test it, learn from it, understand it, communicate with each other and repeat.
Hydar predicts: “As the use of AI-generated content increases, social media platforms may need to differentiate between human and machine-generated content.
“This could be achieved through labelling or other forms of identification to provide transparency for users and reduce the spread of fake news or misinformation. However, implementing labelling would require significant resources and technological capabilities from social media platforms, and it would be important to ensure that AI generated content is not stigmatised or devalued.
“As the technology landscape evolves, it will be interesting to see how social media platforms adapt to accommodate the possibility of labelling human vs AI-generated content.”
Perhaps our Senior Media Consultant Chris Hatherall said it best: “Any technology which helps provide good copy and saves time is useful. However, it needs a human brain, human strategy and human emotion behind it to actually connect with anyone.”
Do you need support to increase brand awareness? Position your spokespeople as experts? Reach new audiences? Gain share of voice against your competitors? Talk to us today about how we can help you reach and exceed your objectives.
This content was human generated.