3 Misleading AI Metaphors And One To Make It Right

Forbes

20-07-2025

Technology is like grammar, a structural system that enables us to understand and navigate the world ... More in a way we otherwise would not be able to, but also makes it harder for us to understand and navigate that which transcends the system. (Photo by Gideon Mendel/Corbis via Getty Images)
No, AI is not a tool, an amplifier or a mirror. And by using these metaphors when talking about technology, you make it harder for yourself and everyone else to develop a responsible relationship with your surroundings. Here's why your typical AI metaphors are misleading – and how you should think and talk about technology instead.
AI Is Not A Tool For What You Want
When we talk about technology as a tool, we think of ourselves as someone who uses the tool for something. It can be to:
Throughout history, we have viewed technology as tools and weapons that can be used for different things with different purposes. Lately we even coined the term technology for good to emphasize that technology can be used for good and bad, and that we – the developers, regulators, and users – determine whether it's used for one or the other.
In other words, thinking about technology as tools makes us think about ourselves as being in control of why and for what technology is used. But, as more and more people emphasize, technology is never neutral. And not just because it reflects the biases of the people who design and generate the data that feeds technology. Also because it has a bias of its own: A simple yet very powerful bias that no matter what we want, we are better off if we use technology to achieve it than if we don't.
This bias, which is a default bias in all technologies, regardless of why and by whom they were developed, takes hold long before we ask ourselves what we want. Therefore AI is not a tool for what you want. It's a technology that allows you to effectively do or avoid doing certain things – while making you think that's exactly what you want.
AI Is Not An Amplifier Of How You Work
Like all other technologies, AI makes you think you want what technology offers. That is, speed, efficiency, and convenience. But unlike other technologies, AI doesn't hide its default bias. It brings us face to face with its insatiable hunger for data, making us ask ourselves if speed, efficiency, and convenience is really what we want. Is it all we want? Or do we have other needs that cannot be met with technology?
These kinds of questions have led people to regard AI not only as a tool but as an amplifier of who we are and what we do. In a 2024 research paper, former MIT professor, Douglas C. Youvan posits that 'AI operates as an amplifier of human traits and tendencies, rather than a neutral or equalizing force'. And he argues that 'far from leveling the playing field, AI tends to accentuate what is already present within individuals, pushing them further along the paths they are predisposed to – whether these are cognitive, economic, behavioral, or ideological.'
This stands in contrast to the mantra that has been repeated over and over since OpenAI launched ChatGPT, namely that 'AI won't take your job, but a human using AI will.' Yet both Youvan and those who argue that AI users will soon replace non-users in all parts of work and life regard AI as an amplifier of how you work. And it is not. You are not designed for speed, efficiency, and convenience. In fact, you are not designed at all.
You are born to live, learn and grow in step with your surroundings. And reinforcing and subjecting yourself to technology's default bias that whatever you want, you're better off if you use technology than if you don't doesn't help you do that.
AI Is Not A Mirror Of Who You Are
Verity Harding is the director of AI and geopolitics project at Cambridge University. In an interview to Forbes India in 2024, she said:
'In some ways, AI holds a mirror up to us and shows us what we are like, particularly these generative AI technologies that are built on existing human data and language online, from books, scripts and blog posts. Although the companies involved have tweaked the algorithms to try their best to ensure it doesn't bring up the worst sides… of course, anything that can show the best side of us can also show the worst.'
Thinking about AI as a mirror is a bit different than thinking about it is as a tool or an amplifier. But all three metaphors suffer from the same mistake in that they make you think of AI as something you can use to achieve something else – better performance in the first cases and better understanding of yourself and other people in the latter.
The problem is not that these metaphors are reductive. The problem is they reduce your understanding of yourself. Instead of thinking of yourself as someone who asks what you want, how you grow, and who you are, the tech metaphors make you think of yourself as a tech user. That is, someone who stands in front of your surroundings with a purpose to achieve something, rather than someone who is part of your surroundings with a responsibility for your shared future.
Like the kids in the pool, we think and talk about AI as tools we can use for what we want, while we ... More forget the structural system that makes us think this way.
AI Is Like Grammar, Shaping How You Think
Tool, amplifier, and mirror are not just misleading metaphors for AI, they are misleading metaphors for all technologies. Because, as the German philosopher Martin Heidegger put it, they make us 'utterly blind to the essence of technology' – that is, the default bias that makes us think that we are always better off using technology than we are not using it.
But as mentioned above, AI differs from other technologies in that it does not hide this default bias. Unlike bows, cars, and other premodern and modern technologies, and unlike the internet, social media, and other digital technologies, AI allows us to see technology for what it really is. Namely, something that has as much influence on us as we have on it.
We are never just users of technology, we are always also being used to promote the bias towards faster and more efficient technology. And while we were previously blind to this mutual influence, the data harvesting of large AI companies has made us see clearly that it is not only the nature around us that is being exploited in the name of technology and progress. We are being exploited too.
To capture this insight and to avoid reducing ourselves and each other to tech users, we need a new metaphor for how to think and talk about technology. Heidegger used the term 'enframing' (Gestell) to describe how the essence of technology positions us as someone who stands in front of our surroundings with a purpose to achieve something. But I would suggest that an everyday phenomenon and term like 'grammar' can do just as well.
Like grammar, technology is a structural system that enables us to understand and navigate the world in a way we otherwise would not be able to – but which also makes it harder for us to understand and navigate the world differently.
Common to technology and grammar is that:
Thinking and talking about technology as grammar opens up a world of possibilities to discover and discuss new aspects of our relationship with AI – and how it shapes our relationship with everything else.