Learning to calibrate trust
AI has been on my mind (and in the news) rather a lot these past few weeks.
It has almost been hard to avoid it.
Between new legislation, AI detection functionality appearing on platforms, businesses scaling (or folding), and reports of autonomous rogue agents venturing off on their own during testing, it has been one reminder after another.
The whole conversation around AI feels somewhat different today than it did two years ago.
At first, it was mostly centred around the capabilities and potential. What could it do? How quickly could it improve? How much time could it save?
Now, though, that excitement of capability seems to have shifted and the focus is firmly on uncovering the answers to some slightly trickier, and perhaps more fundamental, questions.
Ones that centre more around accountability, traceability and trust.
Not on what AI can do, but on how we should use it. And, perhaps even more intriguingly, what AI reveals about us.
This conversation is only just beginning
AI is changing what we perceive as evidence.
For years, a compelling article implied considered thought. A well-written report suggested expertise. Now, I would argue that is no longer completely true.
We want to understand where things have come from. How it was created. What role AI played in its creation. How much human thought contributed to it along the way.
We no longer trust the finished work at face value.
The first visible signs of this ongoing evolution arrived earlier this week.
Without a whole lot of fanfare, both the European AI Act and the California AI Transparency Act came into effect. Legislation introduced not to limit AI as such, but to mandate greater transparency around when and how it is used.
One sentence that particularly stood out while I was reading up on the EU AI Act was the importance and need for humans to be able to “calibrate their trust”.
Re-read that sentence, and let it sink in. Calibrate trust.
It left me wondering whose trust are we determining: AI or humans.
What does trust look like when creating something is no longer evidence that you yourself thought it?
Does trust increase because we are honest and transparent, or does it actually decrease because we question or stop assuming authenticity in the first place?
Regardless of where one stands philosophically on this topic, we are no longer talking about a future possibility or discussion point that requires additional debate. This topic has already arrived.
A new burden of proof
In many ways, we’ve been here before.
Around 25 years ago, particularly in the academic setting, digital plagiarism detection software was deployed to weed out those who took liberties with other people’s work (be it ideas or entire swathes of text or art) and passed it off as their own.
The technology may change, but the underlying concerns feel vaguely similar.
The reaction to the introduction of AI detection software on one platform has been fascinating.
Some viewed it as an unnecessary attempt to stifle or police creativity, or stigmatise the use of AI. Others welcomed it as an important step towards protecting originality, artistic expression and human authorship.
I wasn’t particularly interested in who was right. What interested me more was how quickly the conversation shifted. The sheer strength of the reaction suggests to me that the debate has actually rather little to do with AI. Rather, it’s more about what we individually value as a human contribution.
For my part, I am still unsure whether this software is determining whether my thought is original, or whether I used AI in any shape or form during the process.
If software is used to help improve grammar, punctuation and spelling, is that detected as AI? If it is instead used to help fast-track research to help me better formulate and challenge my own thinking, does that influence my authenticity?
I start to wonder where the line will be drawn. Because it will be drawn somewhere.
A not-so-certain future
Whether you see all of this as sensible governance, unnecessary bureaucracy, or simply the next stage of technological evolution will likely depend heavily on your own perspective, experience and beliefs.
My own thinking may have evolved over the past year, but I confess to still sitting somewhat on the fence.
I still believe that AI can be a useful tool for good. Just imagine the possibilities for medicine, science, engineering, building our future cities, or helping find solutions to our current global challenges. I also believe that by relying on it as much as we have been, we may be starting to sabotage our own judgement.
Perhaps this new legislation is not the story. Neither is the detection software, or AI itself.
Perhaps they are merely the signs of something bigger.
A society trying to redefine what originality, accountability and trust actually mean when intelligence is no longer solely human. In doing so, we may be rediscovering exactly what we value most.
Regardless of where you sit on the AI spectrum, I suspect that this is only the beginning of the conversation.
When marketing in your organisation feels busy but not effective, you are not alone. There is usually a deeper challenge in how marketing is being understood.
The Intuitive Marketing Coach is written for people trying to move beyond activity, assumptions, and box-ticking, towards something more deliberate, valued, and commercially effective. It explores a more intuitive approach to business. One grounded in better judgement, clearer thinking, and a more considered view of what marketing is actually there to do.
If you’re trying to move beyond activity and towards something more logical, intuitive, and ultimately more effective, subscribe.