Think twice: Five key rules for using AI at work

We’ve entered a fascinating moment in the history of work. For many decades, very little about how, where and why we work had changed, but the COVID years upended several of those areas at once, and now AI is doing it all over again.

The rise of AI in the workplace is supercharging some jobs and undercutting others. It is undoubtably going to affect most areas of our lives. But just because it’s new doesn’t mean we have to let it run wild without imposing some rules.

Now, I’m not talking about sorely needed laws and regulation, in Australia and in AI tools’ countries of origin, to keep the models in check. I mean every worker should define their own personal rules of engagement for how they choose to interact with this powerful new technology.

I engage deeply with AI, using it to revolutionise some aspects of my life and intentionally keeping it away from others. These are my five rules for how to handle AI in the workplace.

1. If you expect me to spend time reading something, I expect you to have spent time creating it. AI can pump out plausible-sounding reports in seconds, creating presentations, summaries, blog posts, LinkedIn boasts, emails and other content. Most of it is unnecessary “workslop”.

Wading through reams of AI-generated content is a real and growing problem in many workplaces. Research from Stanford University found that when people receive obviously AI-generated content from a colleague, about half of them viewed the person who sent it as less creative, capable and reliable. One-third even said it made them less likely to want to work with that colleague again.

That’s precisely how I feel when I’m emailed a report to read that has been conjured up by AI with a few bullet-points. Skip the fluff and stop wasting my time.

2. Treat AI as an overly confident junior, no matter how smart it seems. By now we know that AI models have already sucked up every piece of content that’s ever been written or created (almost all of it without permission or payment to its creators).

On the surface it can appear that it’s more intelligent than you. Part of AI’s allure is that it spits out responses that brim with such confidence you could mistake it for authority.

But AI’s “hallucinations” – an oddly human term coined by AI public relations – occur up to a third of the time depending on the model you use. This means you should treat it like a junior colleague and double-check every “fact” that it tells you.

There’s no shame in using AI for tasks, but be upfront when you do.

3. AI’s output is only as good as its inputs. As with humans, the quality of responses improves when you ask better questions. In technology this is known as prompt engineering, or the art of learning how to guide AI to get the best answers.

This skill takes practice and time to develop, and can be the difference between exceptionally helpful new ways of working or undercooked replies. You should rarely blindly accept the first response AI gives, so give it context, ask follow-ups and push the technology as far as it can to get the best results.

4. Be clear on what you’ve used AI for. Almost everyone is using AI at work, and it functions best when it happens in the open. We all know that your daft colleague didn’t just suddenly discover how to write complex emails with multiple subheadings, so be upfront about your usage.

Australia’s former chief scientist, Dr Alan Finkel, recently launched an initiative called Proudly Human to allow people to certify when something has been created primarily by a human.

I met with him recently to hear more about it, and was impressed by his noble attempt to protect creativity in a tsunami of change. There’s no shame in using AI for tasks, but be upfront when you do.

5. Just because you can, doesn’t mean you should. It’s tempting to outsource almost all your research and reasoning to AI, but please think about it before you do. Some things are still better done with the real, not artificial, intelligence that you were born with.

Communication between humans is what makes us, well, human. Yes, AI can assist us with writing first drafts and cleaning up messy ones, but when it takes over our public posts and communication with colleagues, the long-term impact on thinking and creation could be devastating.

AI is still incredibly new in our lives, and we’re all figuring it out together. It’s a very effective tool, but you really need your own personal rules of engagement to ensure that when you use AI it doesn’t end up using you.

Tim Duggan is author of Work Backwards: The Revolutionary Method to Work Smarter and Live Better. He writes a regular newsletter at timduggan.substack.com.

Get workplace news, advice and perspectives to help make your job work for you. Sign up for our weekly Thank God it’s Monday newsletter.