I started my research career in December 2021, just before generative AI entered the mainstream. I was learning how to moderate interviews, structure discussion guides and turn messy data into coherent stories in a hybrid working world, often without regularly sitting alongside the people I was learning from. Then, less than a year later, ChatGPT arrived.

AI didn’t replace what I was learning, but it changed the context in which I was learning it. As a young researcher finding my footing, that felt both daunting and exciting. I wasn’t just developing research skills; I was developing them alongside a technology that was rapidly reshaping how we work.

In many ways, AI levelled the playing field. Suddenly, everyone was learning. Integrating these tools at the same time my core skills were taking shape has felt like an advantage. However, AI is also incredibly seductive. It reassures you that your thinking is strong. It offers smart phrasing, shortcuts and neatly packaged explanations. It can become a ‘safe place’ to ask what might feel like a silly question when you’re not sitting next to colleagues.

That accessibility is powerful, but it introduces a new tension. I think that the most important part of early development in research is not speed or fluency – it’s finding your voice and building judgement. When a tool can generate structure and confidence on demand, it becomes easier to hide behind polish.

You can produce something that sounds sharp before you’ve fully tested whether the thinking underneath is. For young researchers especially, that can blur the line between clarity and conviction. The debate often centres on whether AI risks eroding the human side of research. In my experience, it has sharpened my awareness of what that human side actually is.

At the same time, AI has unlocked an extraordinary expansion of creative possibility. We’ve always been responsible for bringing insight to life visually and verbally, but the range of how we can do that has widened. I can test different framings for an insight in minutes, explore alternative headlines, pressure-test whether a point actually makes sense, or prototype bespoke visuals that act as shorthand for complex ideas. We’re no longer limited to stock imagery or generic templates.

 
AI lowers the friction of iteration. It makes it easier to move from a rough thought to a clearer expression. While that’s a huge advantage,  it also means polish is no longer the differentiator. When everyone has access to tools that can make something sound articulate or look compelling, what really matters is judgement – knowing which version is right, which story holds, and which ideas deserve to be pushed further.

I’ve found it most powerful to treat AI not as an answer machine, but as a sparring partner. The shift in my development hasn’t been about letting it generate work for me, but about learning how to guide it. The clearer I am about what I’m trying to say, the stronger the output becomes. If my thinking is vague, the response is vague. If my framing is sharp, the response sharpens with it. In that sense, AI is less of a shortcut and more of a mirror. It exposes gaps in logic, tests clarity and forces precision.

Working this way hasn’t replaced critical thinking; it has made me more deliberate about it. It has reinforced the fact that communication and storytelling sit at the heart of our role. Insight doesn’t become valuable simply because it exists – it becomes valuable when it is understood, remembered and acted upon. AI can help elevate how we express ideas, but it cannot decide which tension matters most, which narrative will resonate with a client, or which uncomfortable question needs to be asked.

For me, keeping insight human isn’t about resisting AI. It’s about recognising what the technology can’t supply on its own. It can generate words, structure and even a sense of confidence, but it can’t generate judgement.

As early-career researchers, we have an opportunity not just to adopt these tools, but to shape how they’re used. The real differentiator isn’t access to AI. It’s the ability to think clearly, communicate with intent and use technology to elevate – rather than replace – the human craft at the heart of insight.

I’m looking forward to exploring these ideas further at the upcoming MRS Annual Conference panel on young researchers and AI.

Olivia Radcliffe is research manager at Simpson Carpenter

• Radcliffe is participating in a session at the MRS Annual Conference on 10th March: Thinking in the age of machines: how researchers can keep insight human. The conference programme is on the event website here.