We went behind the headlines to look at what ChatGPT is capable of, how it can help teams be more efficient, and what limitations we might need to be aware of. It’s a big topic, opinions vary, and there’s a ton of misinformation out there so we’ve rounded up the key takeaways for you to cut through the noise. Here’s what our experts had to say…

Steven Zimmerman, PhD, specialist in natural language generation

Stance:
Evolution – it’s a big jump forward, it could be as big as Google.

Why is ChatGPT different to other tech?
What sets it apart is the amount of data it’s trained on, (basically all text that was available on the internet up until autumn 2021). Despite not being trained on data since then it may know more up-to-date information because it also learns from interactions with users.

How trustworthy is ChatGPT?
It’s based on probability. It can be totally wrong – currently, there’s nothing to fact-check the information it’s giving you. If you play around with it enough you’ll definitely find the holes.

Sarah Marshall, Global Executive Director, Distribution & Channel Strategy, Condé Nast

Stance:
It’s an evolution – but also a step change. It could help save a lot of time. It’s entering workflow for more technical roles – rather than say, your fashion journos. But it really is being used and experimented with.

How Condé is reacting?
By having cross-functional conversations and exploring how each department can best utilise ChatGPT’s functionality. For them, it’s primarily a backend efficacy tool that has been trialled by the analytics team to translate coding languages and by SEO managers to generate keyword ideas and automate scripts for Google sheets. It could also be used to auto-translate syndicated articles, make personalised content suggestions for newsletter readers, and assist with predictive tagging (which links similar content together to ensure users are served the most relevant content in relation to their interests and previous consumption).

Biggest takeaways:
There are just some things humans can do better, we’re not going to see job losses as a result, but the tech can be harnessed to create more efficient workflows. And of course, whatever you are using ChatGPT for, it’s crucial there’s human intervention. But the developments are exciting.

David Caswell, content automation expert and Executive Product Manager, BBC

Stance:
It’s a revolution – as big as the internet was in the early 90’s – not triggered by just ChatGPT even though this most recent development has brought generative AI to the masses.

Why it’s blown up:
It’s a general technology, like electricity. It’s very easy to use, both in how users interact with it and the way results are delivered. It’s immediately useful to people and it’s free! This rapid and broad adoption has the power democratise innovation.

How should publishers react?
Encourage your teams to interact and explore.

Tried uses:
Automated copy editing, summarisation of one or multiple articles combined, topic-based news briefings, fact-checking, data cleaning, creation of graphic formats, and the production of audio and video.

Key takeaways:
Audiences will ultimately drive the change. Publishers have the responsibility to be transparent when they’ve used ChatGPT and it’s important to remember it’s not always accurate, right now – a degree of caution is needed.

How impactful this new technology will be for the industry remains to be seen. We will of course keep PPA members updated as developments are ongoing. The PPA Festival agenda (be sure to join us on 25 April) will be going more in-depth into risks and regulation, legal implications and IP, and search and audience growth implications.

PPA members can watch the webinar on demand here.