AI is useful, but imperfect



It annoys me when the use of language is loose and incorrect.

Small errors creep in. Words that are spelled and sound the same are used interchangeably.  The trouble is that the words mean different things.

These homonyms – the technical term – seem to go unnoticed either
because AI is not checked, or the person using AI has no clue themselves that they have used the wrong word.

Maybe I am one of a group of few people who care.  Afterall, isn’t it meaning that is most important?

I don’t think so. If we mistake one word for another in something so basic as everyday language, how can we hope to communicate on the
more important things? Imagine an ad saying ‘Sail today’. It would be confusing if the next words were ‘Get 10% off’, and it happened to be that the store had reduced pricing for towels.

This sloppiness in language usage may extend to all types of
things.  Dinner at ate? Pears of socks on sale. The list goes on. Deer friend. You get the message.

Sure, we all make mistakes. Me included. However, the lack
of oversight of AI is disappointing and worrisome.

Imagine a case of a formal document and you use AI and find
words missing or misused? There can be expensive consequences. 

It is not so implausible as it sounds.  I have seen errors in AI produced work recently even on the web site of a well-known company.

AI produced work must be reviewed for correct word usage. Assuming
people will understand just doesn’t make sense – or cents– you choose.