Artificial intelligence – it’s only as ‘intelligent’ as the people using it

That wonderful tool of the impending apocalypse, generative AI, continues to make a nuisance of itself.

Sure, Chat GPT and its ilk have many valuable, time-saving uses in business such as developing marketing initiatives, summarising documents and answering obscure queries like why does Sharon from HR insist on microwaving a foul-smelling haddock fricassee in the office kitchen?

The technology does, however, still have its limitations and these are more or less decided by the sophistication – or lack of it – of the questions it is asked.

Legal citations

We have, for example, experienced a recent rise in the number of correspondents who argue their cases using AI-generated legalese, sprinkling their missives with phraseology they wouldn’t dream of using in their everyday lives.

These include classics such as ‘nor do I accept the alleged amount of its validity’ and ‘any attempt to enforce this disputed invoice may constitute illegal debt recovery’.

These and many others come complete with legal citations to various acts, rules and regulations and – for all I know – the precise contents and nutritional value of Sharon’s evil fish dishes.

AI can be a wonderful aid – Robyn uses it for rationalising notes and finding detailed research sources, as well as editing the many pictures of her French bulldogs she inflicts on her Instagram followers.

In the hands of the fools and vagabonds some of our clients have to deal with, however, it can often be more trouble than it’s worth.

A study from Harvard Business Review and others in 2025 found that around 40% of employees regularly received rubbish, AI-infused emails and that this so-called ‘workslop’ wasted around two hours of valuable work time.

To add insult to injury, these poorly-written, needlessly verbose mails more often than not came from their own colleagues.

Online fraudsters

Generative AI has also been a boon to phishing merchants and other online fraudsters who use it to iron out the bad grammar and weird tone that was always a warning sign in their correspondence.

It has also triggered a massive rise in the volume of e-comms we’re all getting now, another drain on our time, productivity and collective will to live. According to research published in TechRadar, just 13% of email traffic around the world is written by humans.

None of us has the time, patience or energy to deal with something so pointless as sifting out real emails from bots or, as in our case, engaging with gobbledegook from people who are in the wrong in the first place but think AI will muddy the waters sufficiently for them to get away with it.

These are real productivity problems for businesses small and large.

I wonder what ChatGPT would suggest as a solution.