AI can be a problem but there’s an easy way to regulate it

What’s more embarrassing than being a lawyer busted for using artificial intelligence (AI) to draft your court papers? Being a lawyer that does it a year after so many other lawyers have been busted doing the same thing.

It’s somewhat embarrassing to fall into a trap so many others have widely been reported to fall into before. Yet not only did it happen, it still happens and incidentally slows things down; for all the time saved drafting with AI, courts, lawyers and admin had to spend time investigating the use of AI and finding the outcome.

Similarly, management reports, university assignments – any document that people rely on – will take so much more time to confirm whether it was written by a human or a computer. Sometimes, it doesn’t matter. Other times, when there’s some money or certification on the line, it absolutely does matter.

Some might not care about whether certain documents they’re receiving are conceived from AI. I use a great AI tool to keep minutes of all my meetings. That’s hardly contentious and frees up so much time to do actual work.

In other instances, they may well care. Take, for example, when paying legal fees only to find out your lawyer is referencing non-existent case law on the advice of some software. Once you get over the hilarity, you’ll soon come to realise that in all instances of AI use, the person making use of the software is rarely the person who is going to be the beneficiary or end user of the product.

ALSO READ: Is South Africa ready to handle AI?

Typically, we don’t trust AI completely yet when it’s our own interests on the line?

Even when it comes to assignments, it’s just having something on paper that’s helpful and how would some marker who has to go through 600 of the things bother to check yours? Even if they wanted to, there’s so much security in knowing that there is no real way that they can.

Moreover, there’s also increasing safety in knowing that even if there were to be some kind of AI regulation, there’d be limited ways of enforcing it. You’ll get no love trying to force some complex set of rules on the data companies running the things and you’ll never be able to force every user to comply with all the laws.

So make regulations on the AI generators incredibly easy and allow the public to vet their own documents. In other words, compel AI generators to have a public database of all documents generated so that anybody can check a document before them against the database.

ALSO READ: From plagiarism to AI: Protecting creativity in a digital world

It may sound simplistic but that’s because it is… as well as effective.

Remove the shadows which anybody can hide behind claiming to have drafted the document and allow people to confirm for themselves whether a document was AI generated or not. If nobody cares, then the system won’t be burdened with overregulation but in instances where people do care, they can check.

Best yet, there’s no overlord dictating what people should or should not care about.

Proper regulation will be to the benefit of the public and should protect the public from abusive practices. Charging them money for documents that are artificially generated or unleashing clueless idiots into the world who used AI to get their way through a degree or course is something we should guard against. It also doesn’t seem to be overly cumbersome to require a public database of previous results and the search parameters that lead to them.

We could just leave it and avoid regulation altogether but there’s a reason why a lawyer thought it was okay to generate their court papers artificially despite some busts a year ago; clearly they’ve been getting away with it, as have many others and they just happened to be caught. But the risk/reward clearly seemed worth it.

So if the public is to be protected, we need to do something and empowering the public to check if the document in front of them is AI generated is the way to go.

Content generated by ChatChemaly.