Artificial Intelligence is supposed to help us but the lack of reasoning it engenders means future generations will be less intelligent.
Use it or lose it. This certainly applies to languages, and to everything else. The recent news that tomorrow’s generation of children will be less intelligent than their parents’ generation is alarming and brings several issues into play.
Firstly, this is nothing new. Even today with our modern technology we do not understand how Ancient Romans managed some of their engineering feats and those of us who read ancient texts discover that thousands of years ago, people were far more advanced than we might imagine, building bridges across wide fast-flowing rivers (London), having names for planets which were supposed to be discovered in the 18th and 19th centuries (Babylon), marvelling at water clocks chiming the hour (Rome) for instance. Examine the quality of Roman mosaics from the 1st century BC and compare them with those of the 3rd century AD and you see a clear deterioration.
Secondly, despite our advances in technology, the latest research reveals alarming trends. Causes and solutions?
It is obvious that if a child stops doing mental arithmetic and arrives at an answer by pushing a button, it will not make sense to spend time reasoning when (s)he can find the answer to a complex arithmetic problem in a few seconds. On one hand, (s)he has saved time, on the other hand, (s)he has stopped using cognitive chains. The reasoning process has been surpassed, like a ball lobbed over the top of midfield in soccer straight from goalkeeper to forward. Pass. Goal. Who needs midfield players costing hundreds of millions?
It is also obvious that if a child is presented with multiple choice tests, circling ABCD with a pen and not describing how (s)he got to the answer in writing, then the same cognitive process will be lost, let alone spelling skills, as automatic correctors not only do that job but also suggest the rest of the line. Or if using Artificial Intelligence, the entire text. Or doctoral thesis, complete with sources.
Now let us go fast forward to the world of music or the visual arts. When I was involved in songwriting (before the Internet appeared, I wrote some 50 songs and published at least 40 of them) we used to spend a day in a recording studio recording a song, placing the percussion on the lower tracks (1,2), then the bass on 3,4 and then the other instruments on tracks 5,6 and finally, the voice on the higher tracks (7,8 of eight, 15,16 of 16 or 23,24 of 24 and so on). You could improve the quality to some extent by turning buttons and sliding sliders up and down (contrary to popular belief that is the sound engineer’s job, not the authors, composers or performers of the songs and that is not what people do in a recording studio. They record).
These days, you can record a song on a laptop. You record the music, then you sing over it, once. You see the voice as a soundwave and can tweak it on the screen. You can make an old lady sound like a robot, you can make a soprano sound like a bass, you can polish in three minutes what used to take a day to do.
But worse. You can even press a button and get a song, complete with lyrics, asking AI. Write me a song with 4 verses and a refrain/chorus with the theme Blue Moon Clouds. Style? Hip Hop. Suggested rhythm. No, faster. Stronger snare. Add bass guitar. Trombone here. Violins there. That’s it. Done.
Trouble is, who are the writers, and composers? How do you calculate the copyright, because the AI has lifted it from somewhere, probably from many different sources but there is a chain of notes from somewhere that it has copied. This area, that of intellectual property, is of great importance to those who are having their work copied, for free (in the old days you used to make a dollar a disc but with streaming these days that has gone). It also means that if we do not use our skills to build cognitive chains, we will lose the ability to make music or create visual arts.
Thirdly, watch a young person these days and count how many seconds (s)he can sit still and concentrate. Unless of course (s)he has six different machines flashing their screens besides the television, which anyway is considered “boring” and is merely a source to get the Internet on a wider screen. Attention spans are measured in seconds and with this, the ability to sit and read a book and to learn how to think and reason. What happens if the push-button solution is not available?
My mum once went into the bakery and asked for 100 bread rolls at, let’s say, ten cents each (it was not in the USA). The calculator was not working. First girl: “one dollar”. Second girl: “100 dollars”. The manager “One dollar twenty”. My mother put the ten dollars on the counter, said “Organise yourselves”, took the bread and walked out.
Let us take this debate one stage further. If people lose the ability to reason, they will turn into gullible imbeciles. In the past, if a politician claimed that (s)he would provide homes for everyone, free education, free utilities, free healthcare, in the western world at least, people would say “Yes but how do you pay for it?” (In the socialist countries they managed perfectly well but that is another story). People would argue that you either increase taxes, or borrow from the markets. One way or another they would hold the politician accountable through reasoning.
And today? All it takes to win an election is a sexy soundbite. “Make America Great Again” (whatever that means). “Get Brexit Done” (How to do something that cannot possibly work, by definition). So far-right populist groups are firing soundbites at populations, whose younger members are falling for it. “Clean up our politics”. Or on a more xenophobic slant, “The immigrants are all coming here to take your jobs”. “Hospital and school places are being taken”.
For copyright and intellectual property rights, there cannot be a legal solution without a legal framework so cyber law has to be written down and enforceable. As we see with international law, writing it is one thing, implementing it another. Those sitting on high moral horses today forget what they did in Iraq and Libya, with zero context, because they managed to get away with it and ride the wave of uproar, which soon died down. In the world of music, possibly, the rule in the old days of considering as plagiarism six consecutive notes and a finite number of words in lyrics could still apply to AI?
In a wider context, things have to begin, again, with a framework and the framework of a society begins with education. First question, Are our classrooms today adapted to the socialisation of those who sit in them? In other words, Is the classroom socialised for the pupil? Or are they fundamentally the same as they were in my time (born 1958)?
Next, the content of the curriculum has to offer solutions. No calculators in classrooms until cognitive chain-building has been established because with a developing brain, either you acquire the skills at the right time or you never manage to acquire them. The experts might agree on age twelve, for example. Multiple choice tests, OK but also space for reasoning, in writing. And other information, for instance what exactly am I voting for in the next European Parliamentary Election? How do I start a company? How does the Government work? When I come out of school, am I fully prepared for citizenship?
As we see it is about asking questions and reasoning out the answers, starting at the beginning and using common sense, something which AI cannot replace and never will. Watch this space. If we do not take action, we may as well change the name of the species to Homo stultus and remove the word sapiens.
After all, we begin to know less and less.
Timothy Bancroft-Hinchey can be contacted at Timothy.hinchey@gmail.com