AI – Intelligence That Can’t Think
That don’t mean it’s cheap!
Intelligence is defined as the ability to acquire and apply knowledge and skills; which AI can do. Thinking is not the same. The Oxford definition is ‘The process of using one’s mind to consider or reason about something’. The question of whether AI actually ‘thinks’ is far from decided and frankly it doesn’t look good for AI; not even its strongest proponents will claim the computers are thinking, they claim that “the day is coming when it WILL think!”: A claim without proof. It is that ‘using one’s mind’ part that we will consider here.
How does the brain work? The truth is we don’t know. I have spent my life in active research on this question – it’s called teaching. The bane of every teacher’s existence is the highly paid ‘expert’ who is hired each year to explain to all the teachers, many with more years of practical experience than the ‘expert’ has been alive, how to teach kids faster, easier and with less paper! I finally concluded I would never be asked to listen to an expert on how to teach kids better! That’s because all those experts were sitting next to me, bored out of their skulls.
That is until AI. NOW we have an expert that cannot think, but sounds really smart, and those who love experts know that sounding smart is the most important part of the job description. What about that thinking part? Well AI cannot think but fakes it pretty good – until the hallucinations. That’s what designers call it when AI comes up with responses that are obviously false. The truth is that hallucinating is very little different from ‘intelligence’ if the goal is to sound like you know what you’re talking about, but can be very different than thinking, which the designers readily admit AI cannot do yet, but which they would love to convince us is certain to happen at some point in the future.
If we don’t know what thinking is or how the human brain accomplishes this miracle, how will we know when AI is thinking?
WE WON’T!
The argument can be made, and will be here, that the methodology by which the brain causes tissue to react and create an idea can never be understood by the same human brain that creates the idea. In mathematics this question has already been addressed by Rene DesCartes in Cogito ergo sum! I think therefore I am!
There’s that ‘think’ word again; which is defined as ‘using one’s mind’; which is defined as the part of our psyche where ‘thinking’ occurs. The ideal ‘circular reasoning’ of logic; none of these enlighten us a whit regarding what thinking is or how it’s done. The concept that this creation could ever think must also face the truth that, by design, the only thing AI can do is imitate thinking: How? By observing humans who are thinking. Whenever you believe AI is showing ‘smart’, know that it is imitating the smart humans on which it was trained.
Whatever the abilities of AI are, the one thing they are not is cheap! The massive data centers required are a part of the story being quite efficiently swept under the rug. The demands made on the energy grid and water supply are equally massive, and we are just ignoring the additional costs incurred relative to the increased demands on these limited resources.
Marketing strategy says create demand for your product. For AI that mandates selling the belief that AI can THINK – better than humans! Naturally, that AI cannot do such thing presents a problem, but now they’re selling the idea that time will solve this problem, maybe not now, but soon! How firmly someone promotes this provides a pretty clear correlation to how much money they hope to make from AI.
Perhaps the more pertinent and instructive area to investigate is where the AI enthusiast sees money to be made: If you can’t see public education as a potential gold mine you have no imagination. Replacing workers whose efforts require little actual ‘thinking’ is well suited to AI. All that is required is to convince the public that teachers fall into that category: Now the ulterior motive behind the belittlement of teachers and the work they do becomes clearer. Next is to make the cost of public education as prohibitive as possible, a strategy for which vouchers are tailor made. Once the cost of education is steep enough it doesn’t matter whether you agree with AI teachers or not, it’s all that the budget allows!
Naturally nobody wants their kid taught by AI: That’s the beauty of vouchers! The myth on which vouchers are sold is that it will allow anyone to get out of public schools, when the reality is the private schools that the voucher can pay for provide even less education than the public schools from which the students supposedly escaped. The wealthy who would never dream of allowing their offspring to go to school with those children get a nice deduction on the cost of the excellent private schools they attend; and at the same time the public schools must be run on whatever public funds are left when the vouchers are paid – meaning all they can afford are AI teachers. Even better it may be possible to program the AI teachers to teach a ‘preferred’ curriculum; something you damn sure can’t force on human teachers who have those annoying consciences.
Scary scenario? I’m scared shitless! Can this be prevented? On the one hand public education is so popular I have serious doubts about the likelihood of eliminating it. On the other hand removing funding has been shown highly effective in decreasing the relevance of a service: Vouchers have an excellent chance of permanently changing how we view public education, and not for the better.
Our view of education – our view of what ‘education’ means is in serious need of total revision. AI sits squarely in the middle of this, the more we learn about ‘thinking’ the clearer the difference between ‘intelligence’ and ‘thinking ability’ will become. The artistry and instincts that make teaching the most exquisite example of ‘using one’s mind’ should be at the heart of education. Will AI be useful to teachers? Absolutely! Can AI replace teachers? NO!
The solution to this is as simple as it is difficult – we have to let teachers control how students are taught. Not as individuals but teachers must be given control of their own profession – if you have any doubts as to who controls how teachers teach just ask one. The difficulty lies in the fact that it won’t be cheap – a fact we seem to accept for other professions!
Local control of education is an American tradition that the federal government has been working to usurp for at least four decades, beginning with Ronald Reagan’s touting of our ‘failing’ public education which the GOP – with ample Democratic help – has been turning into a self-fulfilling prophecy ever since. The feds used constitutional rights to ensure schools provided the minimums all citizens were entitled to, then used the bludgeon of threatening to withdraw the funds that enabled those minimums if schools did not teach the core curriculum and pass the required tests to please those experts in education; the politicians! The primary reason I retired from teaching, a job I could have done to my grave, was because the politicians began dictating what and how I could teach.
The federal money is desperately needed and irreplaceable, but the strings the politicians insist on attaching must be cut! Fortunately federal zeal has attached so much to its funding – look closely at the cost of standardized testing – that already some school systems have elected to cut the federal control and give up the funds, and actually lower their costs! Attempting to implement AI teachers will involve substantial initial costs and may help stimulate the drive to get the feds out of our public schools, where frankly they never should have been in the first place; other than ensuring the Constitution is followed.
The way is clear but the road is difficult. History shows that education is a requirement for civilization; and this country was built on the power of opening education to everyone! It is why our universities are the envy of the world. This millennium we appear to be doing our level best to reverse that legacy for an imagined improvement of the bottom lines of people who specialize in reaping benefits from the work of others. History will not be very forgiving of our stupidity.