Too often discussion around using AI is “either/or” — an assumption that you either use AI for a task, or do it yourself. But there’s another option: do both.

“Parallel prompting“* is the term I use for this: while you perform a task manually, you also get the AI to perform the same task algorithmically.

For example, you might brainstorm ideas for a story while asking ChatGPT to do the same. Or you might look for potential leads in a company report — and upload it to NotebookLM to perform the same task. You might draft an FOI request but get Claude to draft one too, or get Copilot to rewrite the intro to a story while you attempt the same thing.

Then you compare the results.

The best of both worlds

There are a number of advantages to parallel prompting:

• You get the ‘best of both worlds’: AI can draw on a vast amount of indiscriminate information; you draw on a narrower but more selective knowledge base specific to your context
• It reinforces and expands knowledge: drafting a prompt encourages you to ‘teach’ the AI what you know, a common way to reinforce knowledge. It might also encourage you to learn more about something in order to design a better prompt.
• It forces externalisation and review: drafting a prompt also encourages you to break down a problem and consider your criteria for a successful result. Comparing AI responses with your own results encourages you to be more critical about both.
• Human bias and AI bias can both be addressed: AI is biased in different ways to humans, who have cognitive biases. Comparing results allows you to address both.
• It improves prompt design skills: the more prompts you design, the better you get — you’re skilling up, not deskilling.

Of course, parallel prompting means that you’re not necessarily saving time by getting AI to do a task for you — at least in the short term.

But it can save time in the long term.

For example, an FOI request drafted this way may lead to fewer problems further down the line. Or interview questions generated using parallel prompting might save time chasing up extra information later. You might do some data analysis at the same time as asking an AI tool to perform the same analysis — highlighting mistakes that you (or AI) tend to make, or better methods you can use in future.

In other words, it’s a learning process through being exposed to new ideas and methods — precisely the opposite of the deskilling that can result from poor use of AI.

*Parallel prompting has also been used to refer to using multiple prompts with different personas at the same time. That’s cool too, but not how I’m using the term.