Staying Human at the Speed of Technology
Seventy years from now
This week I was lucky enough to spend an evening at New Scientist's 70th anniversary event, and I've been chewing on it ever since.
The question on the table was a big one: what could the next 70 years look like? We heard from Microsoft's technical fellow and chief scientist, the Director-General of CERN, the Director of Science at the Science Museum Group, and the UK's Minister for AI, supported by a panel that pulled the threads together. Topics ranged from particle physics, to turning fatal diseases into simply conditions we live with, to satellites and the AI tools many of us open before breakfast.
Enthralling topics aside, the arc for me was balance. How do we keep pace with technical evolution without losing what makes us human? It's an abstract question, a personal one -- and a leadership one. That's probably why New Scientist used the night to launch New Scientist Business, a new offer to support leaders as science becomes a bigger catalyst for where business goes next.
Three things that stuck with me
AI adoption isn't the same as AI value. One conversation landed pretty close to home. Plenty of organisations have "adopted" AI, but people still aren't getting the right answers or the right output. Access was never the hard part, right? Knowing how to work with it is - a human skill in its own right.
Someone always asks: What is it for? There's a long-running public debate about the money spent on CERN - and on big technical advancement more broadly. Why fund something whose payoff you can't see yet, and perhaps don't understand? It's the same tension every leader feels when investing in capability before the business case is fully proven.
One day there could be more satellites in the sky than stars we can see. I haven't stopped thinking about that one. It's a striking image of… progress, sure… but what we might trade away in this race without noticing?
Meanwhile, the industry is warning us to slow down
The timing was hard to ignore! Sam Altman recently told Fortune that OpenAI won't go public in 2026, pointing to AI safety concerns. He also suggested the leading labs may be close to agreeing to slow development and work together on safety. The same weekend, Anthropic's CEO Dario Amodei published an essay calling for the pace of capability improvement to slow.
When the people building the technology are publicly debating the brakes, it's a signal we need to take account of.
One moment from the evening called this out. More people said they'd trust advancements to come from academic institutions than from big business. Only 5% of us trusted big business. (Wonder who responded with that?!?). The future is a group effort: academia setting the evidence and the guardrails, government enforcing them, business putting them into practice, and each of us choosing how we use these advancements day to day. Wherever you sit on that spectrum, it’s indeed an “us” problem, not a “they” problem.
Two things can be true
Talking these topics over with my husband later that evening, we realised that many frame these issues as a choice; and so we all get stuck.
Part of humanity is to seek understanding, we should push science forward and fund the CERNs of the world, and as humans, we should keep investing in and protecting humanity right now.
We can totally raise an eyebrow at the timing of OpenAI's warning alongside its investment plans, and still take AI's risks seriously - and importantly, act on them.
AI has real potential to drive innovation, and we need to understand the human and environmental cost each time we prompt.
Polarities are super seductive, simply because they're simple. All in or opt out? Evangelist or sceptic? Nothing is black and white; and sorry, George, we don't get to live at either end. As humans, as leaders and as a society, we have to hold both and make deliberate choices in the middle.
Intentional, not automatic
How do I deal with this personally? Intentional use. Using tech to help me think, not do my thinking for me. Our cognitive skills, our individuality, our experiences, our lateral thinking, reasoning - our humanity - are the assets here. The way you use AI should support these, not replace them.
For sure, that’s harder than it sounds. As humans we like to optimise and make things frictionless, (I certainly do), but some friction really is worth holding on to. I recently came across Stop the AI Drift, a campaign from the Center for Humane Technology, one of many conversations happening right now about intentional use. One of its suggestions is an AI-free day each week, which I realised I already do often; it keeps me connected to my own discernment.
On a practical level, Anthropic's AI Fluency framework is a useful lens. It breaks working with AI into four Ds
Delegation - Deciding what to hand to AI and what to keep. “Is this a task where my judgement is the point?”
Description - Explaining the task, context and standard clearly. “Would a new colleague understand this brief?”
Discernment - Evaluating what comes back. “Is this right, or does it just sound right?”
Diligence - Owning the outcome and using AI responsibly. “Am I comfortable putting my name to this?”
Intentional use also means choosing the right tool for the job. Not everything needs an AI agent. Really, it doesn't. A predictable, repeatable process is often better served by a plain automation, with agents saved for work that needs judgement across messy inputs.
Sometimes the intentional choice is no technology at all. With so many tools to take away the friction, I've been drawn back to pen and paper for thinking time, and I'm not the only one. Staying grounded can be as simple as protecting the space where you think before you prompt.
The output is an artifact, not the answer
What AI gives you is an artifact: a draft, a starting point, something for you to review. Treating it as the finished article is not only lazy, but dangerous. Especially if actions are taken based on what is output. (Though, why would you output anything that isn’t used?)
It's also the core of the adoption issue, in my opinion. When people say AI "isn't giving the right answers", the gap is often on our side of the screen, rather than the tech itself. Briefing well is a skill. There's a wealth of material out there on prompt engineering, so skill yourself up. A good rule of thumb: brief AI the way you'd brief someone you're managing or teaching. Spend time in a notepad/word document really getting it right.
Reviewing well is also a skill. Check the facts in front of you, whether they came from AI or from Steve in Business Development, test the logic against what you know about your business. Ask what's missing, not just what's wrong.
Staying human in the meantime
Of course, no-one in that room could tell us for sure what the world will look like in 70 years. But I left fairly sure of one thing: the people who'll navigate it best won't be the ones who picked a side. They'll be the ones who hold on to both.
