Settling into an AI-assisted workplace

Kit Allen 20th of September 2023

Six months ago, we published a short essay on how AI is changing the tech workplace. It being National Coding Week and all, this seems like an appropriate time to revisit it.

1999 called, it wants its cliche AI generated video content back

At the time, we postulated that there’s one of two ways to approach this looming spectre – either to run around screaming with our collective hair on fire, or put the new tools to work for us. Adapt or die, as the phrase goes.

While six months might seem quite a short timescale to stop and look back, it’s actually quite notable how much has changed. From the wider industry, there’s a distinct bubble of new start-ups applying AI to everything from video production (see above) to full-blown, complete website design and development.

Meanwhile, larger scale companies are scrabbling to keep up. Google is pushing Bard for domestic users and Duet for workspace owners as digital assistants, as well as championing sector-specific tools such as Med PaLM 2 (which is outscoring human doctors in the medical sector). GitHub’s CoPilot and StackOverflow AI Search are pulling the technology into long-beloved platforms for tech knowledge sharing. Even if you don’t realise it, it’s kind of everywhere right now.

So, where does it fit in at Creatomatic?

Well, we’re leveraging Bard and ChatGPT a lot, for starters. ChatGPT’s not perfect for copy yet – it definitely has it’s own tone of voice which is a bit more ‘breathlessly-enthusiastic-and-sincere-and-let’s-use-all-the-hashtags-too’ than we’d personally like. (Pro-tip – look for five paragraphs of roughly equal length, with the final one starting ‘In conclusion’). But it’s fantastic for replacing ‘lorum ipsum’ copy in mock-ups or product drafts, as well as providing a starting point for helping our clients draft an outline for copy.

Over and above this rather prosaic use, ChatGPT is super-handy for repetitive or time-consuming code tasks. For example: this week, I had a job to take a data source and convert it to HTML for display as part of a function. ChatGPT automated this job – which would have taken a good few hours to map out manually – in a few seconds, leaving me with a beautifully formatted code snippet to edit as required.

Last time we looked at this, I noted that we hadn’t used it much in design – yet. But that’s moved on too – Adobe’s Generative AI was launched over the summer and provides an impressive range of tools for image manipulation and generation. We haven’t quite reached DragGAN levels of manipulation yet, but it’s not far over the horizon.

In conclusion*, AI’s insinuating its way into the tech sector pretty much everywhere we look – but we’re enjoying using these new tools and looking forward to the next leaps forward too. And the robots don’t appear to have taken our jobs yet either.

*Chat GPT didn’t actually write any of this. Just making sure you were paying attention there.