I recently came across a remark by the American poet Maxwell Bodenheim in the quotations library I have been building for V2Melody:

“They rarely ever judge a thing merely on the ground of whether it has achieved that which it obviously set out to achieve.”

He wrote that in 1919, long before artificial intelligence, social media, or the curious modern custom of announcing one's opinion to complete strangers.

It struck me as remarkably current.

Until recently, V2Melody existed mainly in its own quiet corner of the internet. People arrived because they were looking for a poem, a poet, a quotation or perhaps something musical. They wandered around, listened to something, and went away again.

This suited me perfectly.

Recently, however, I decided that if I wanted more people to discover the site, I ought occasionally to mention its existence elsewhere. So I ventured onto social media.

This is rather like opening the garden gate and discovering that one has inadvertently provided access to the motorway.

It didn't take long.

One "gentleman" took a look at V2Melody and proclaimed: “Yikes what is this slopper horror”.

I replied.

He blocked me.

This was an admirably efficient resolution to the disagreement, and one which saved me the trouble of blocking him myself.

But the encounter made me think about Bodenheim.

There are perfectly reasonable arguments to be had about artificial intelligence. There are questions about copyright, employment, authenticity, artistic value and the effects of increasingly capable machines upon creative professions. I find many of those questions fascinating.

“Yikes what is this slopper horror”, however, is not an argument.

It is barely even a coherent sentence, although I suppose it does have the considerable virtue of telling us rather more about the critic than the thing being criticised.

Bodenheim's test seems much more useful.

What was the work trying to achieve, and did it achieve it?

V2Melody does not claim that its musical interpretations were performed by human musicians. It does not claim that its images were painted by human artists. I am entirely open about the use of artificial intelligence because the project could not exist on anything approaching its present scale without it.

Its purpose is something quite different.

I want people to encounter poetry.

I want someone who has never read a particular poet to hear one of their poems and become curious about them. I want someone to discover a forgotten poem because an image caught their attention. I want words written a hundred, two hundred or five hundred years ago to acquire, for a few minutes, the immediacy of something created yesterday.

That is what I set out to achieve.

Whether I succeed is a perfectly legitimate subject for criticism.

Does the music illuminate the poem or smother it?

Does an image capture something essential about the words or merely decorate the page?

Does the experience encourage someone to read another poem?

Those are interesting questions.

“AI made it, therefore it is rubbish” isn't criticism of the result at all. It is criticism of the tool.

It is rather like reviewing a novel by objecting to the word processor.

There is another aspect of social media that fascinates me.

I am quite a large man. When I walk along the street, strangers very rarely approach me to announce that I am walking too slowly, occupying too much pavement, or generally conducting myself in a manner that has disappointed them.

People are remarkably civilised on pavements.

Put those same people inside motor cars, however, and an extraordinary transformation can occur.

The mild-mannered accountant who would apologise profusely for brushing against your sleeve in a supermarket suddenly becomes, behind a windscreen, quite prepared to provide a detailed assessment of your intelligence, ancestry and probable destination in the afterlife.

The windscreen changes things.

There is glass between us. There are locks on the doors. Most importantly, there is an accelerator.

Social media merely replaces one screen with another.

The windscreen becomes the computer screen or the phone screen, but the principle is much the same. Distance provides courage of a peculiar sort. People will type things to complete strangers that I strongly suspect they would never say while standing three feet away from them in a café.

They hide behind screens, whether those screens are made of glass or pixels.

Perhaps that is why I find the block button such a wonderfully civilised invention.

It is the internet equivalent of compelling the disagreeable chap shouting abuse from behind his car windscreen to drive into a tree.

I have no desire to persuade everyone to like artificial intelligence.

Nor do I need everyone to like V2Melody.

I don't particularly want everyone to like every musical interpretation either. Taste would become extraordinarily dull if we all agreed.

But there is another problem with dismissing V2Melody simply as “AI-generated”.

It leaves out half the equation.

Fortunately, my long-suffering wife, Rocío, has solved this problem.

Today she came up with a new formula for V2Melody:

NI + AI = V2M

Natural Intelligence + Artificial Intelligence = Verse to Melody.

I rather like it.

It also happens to describe the project much better than the increasingly meaningless label “AI-generated”.

Artificial intelligence can generate music, create an image, analyse a poem, suggest possibilities and perform tasks at a speed that would have been unimaginable only a few years ago.

But it doesn't decide which poem I shall work on next.

It doesn't decide what the poem means to me.

It doesn't choose the musical style I think might bring those particular words to life.

It doesn't decide which interpretation deserves to survive the many I reject.

It doesn't decide whether an image captures the emotional heart of a poem.

And it certainly hasn't spent its retirement reading tens of thousands of poems simply because it loves poetry.

That part is mine.

Natural intelligence supplies the intention, curiosity, taste and judgement. Artificial intelligence supplies capabilities I do not possess and could never afford to employ on the scale V2Melody requires.

Neither part embarrasses me.

Indeed, the combination is precisely what fascinates me.

Human beings have always extended themselves through tools. A camera allows someone who cannot paint to preserve an image. A printing press allows one person's words to reach thousands. A synthesiser produces sounds that no violin or piano can make. None of these things has an intention of its own.

Someone still has to decide what to do with it.

Put natural intelligence and artificial intelligence together and something can emerge that neither could have produced independently.

NI + AI = V2M.

Thank you, Rocío.

I shall pretend I thought of it.

Which brings me back to Maxwell Bodenheim.

Before dismissing something because we dislike the means by which it was created, perhaps we should first ask what its creator was trying to do.

Then look at the thing itself.

Did it succeed?

Did it fail?

Did it move you, interest you, irritate you, surprise you, teach you something or make you curious enough to look further?

Those responses tell us something about the work.

The label on the tool tells us very little.

As for the occasional vacuous stranger who arrives through social media simply to tell me that everything I have made is dreadful, I'll point him towards the nearest tree or bottomless ravine.

Naturally, I'll do so from the other side of my screen, while his opinions can continue to fill the much-needed empty spaces of the online world.