Would You Trust Your Florist to Fix Your Brakes?
Almost nobody would say yes to that question. It feels obviously wrong before you even think about why. But sit with it for a second, because the reason it’s wrong is more interesting than it first appears, and it has a lot to say about what’s happening right now with AI.
A mechanic is licensed. There’s a legal wall between “person who arranges flowers” and “person who is allowed to touch your brakes,” and that wall exists whether or not the florist feels confident with a wrench.
Those protections aren’t in place for all industries. Technology, for example. Software, app and web development. Cyber security, design and branding. Cloud services. None of these have ever required a licence, but what kept them mostly in check wasn’t the law, it was a healthy amount of friction.
A safety barrier was created by the fact that if you didn’t know what you were doing, it was glaringly obvious. That friction is mostly gone now. AI will hand you something that looks finished, reads confidently, and runs, right up until the moment it doesn’t. The gap between looking like you know what you’re doing and actually knowing, has never been thinner, or harder to spot.
Now you might think, what’s the harm? I’m not saying all these fields hold the same level of danger as dodgy brakes, but they still have the potential to wreak havoc on your business.
In any given week you’ll find someone building a website who doesn’t know what DNS records or caching are. A “designer” who can’t produce a vector file, the single most basic requirement of the job. Someone spitting out unoriginal content that’s practically begging to be slapped with a copyright infringement claim. Someone who was in a completely different industry six months ago, now running a tech company.
People have always worked in fields they weren’t trained for. What’s new is that they can now do it with minimal friction. They don’t even know enough to know what they don’t know, and that’s where the danger lies! Someone can build and sell a service using a tool that writes code they can’t read, touching systems they don’t understand, with none of the experience to know when a mistake is serious enough to matter.
All this raises some genuinely uncomfortable questions. We live in a society built around intellectual property rights and liability. If a professional makes a mistake on client work, they’re liable for it. There’s now consultants calling themselves AI experts, charging to set up systems they don’t fully understand. When someone implements a system designed and built by AI, or follows its instructions to build one themselves, and it fails, creates a privacy breach, or gets hacked, who’s liable? Don’t ask me, I’m not a lawyer.
If you’re a contractor using these tools, does your indemnity insurance cover you for these systems failing? Worth checking. The speed at which this has unfolded means the correct checks and balances and regulations are not in place.
None of this means AI is bad, or that beginners shouldn’t use it, or that everyone working outside their lane is reckless. But I think some caution is warranted, and that AI doesn’t yet replace humans with the appropriate experience to know the risks, or what your prompt failed to capture.