A solution in search of a problem

“Writing is the most fun you can have by yourself.” – Terry Pratchett

Back when I worked in the IT department of a higher education institution, many was the time a member of the university executive or administration would come into the IT offices with a brilliant new idea.

This was usually something like “We need a blog!” or “We need a wiki!” back in the old Web 2.0 days when blogs and wikis were still cutting edge web technologies. When asked what we needed them for, the answer was more often than not along the lines of “Well, Warwick have got one!”

In other words it was just a case of keeping up with the institutional Joneses. One of the IT management team had a very astute response to this kind of request which I’ll repeat here.

“What do you need to do? Don’t come to me with a solution. Come to me with a problem.”

It’s a piece of advice that I’ve started to think of recently as software houses and companies scramble to add AI to their products and services, chiefly because other people are doing it.

closeup of typewriter ribbin in a typewriter upon which someone has just typed 'something worth reading'

In many ways AI – or at least GenAI – is a solution in search of a problem.

AI technology itself is a very powerful and useful tool. It has any number of applications where it could advance human knowledge, for example in medical science (where it could help save lives) or planetary astronomy (where it could help discover life). There are doubtless many more.

It also has the potential for streamlining repetitive and monotonous administrative tasks like analysing large datasets or archive entry labelling, leaving people time for more challenging work that requires thought and analysis. This more challenging work includes writing.

Yes, even writing emails.

Writing emails is not tedious admin. It helps marshal your thoughts. As a developer, more than once I have started an email to someone asking a technical question only to come up with the answer myself as a result of having to shape my query into something I’m sure the other person will understand.

Writing trains your brain and develops analytical and critical thinking skills. Using GenAI to replace this is counter-productive. GenAI text is all surface and no substance, because the process of writing is everything.

A useful analogy would be weightlifting. Sure, you could invent a machine to move the dumbbells up and down for you but it wouldn’t do your muscles any good.

Promoting GenAI for writing is encouraging mental idleness. What is worrying is how much effort is being put into this, rather than – for example – encouraging maximum accessibility in software and online. If people can spend time writing a prompt for GenAI then they could certainly spend time writing alt text for an online image or thinking about how an experience could be made universal for everyone no matter how they're accessing the content.

Unfortunately, more often than not, accessibility testing is an afterthought whilst incorporating GenAI is front and centre.

Steve Krug’s famous quote “Don’t make me think” referred to the end users of systems, not the people designing the software or writing the documentation. We should all be thinking.

The way that companies keep replacing human copywriters with GenAI is stating to feel like the quest for the perpetual motion machine. The creativity has got to come from somewhere. No matter how many springs and wheels you add, without external input it's ultimately doomed to mediocrity and inanity.

Comments

Popular posts from this blog

The Weird Case of Weirdcase

Neurodiversity in the Workplace

Top of the Cheats