Who watches the AI ‘fact’ generators? 

 

Today, our betters in Westminster have lovely, cheerful ‘news’ for us:“Newspaper headlines: Energy bills to ‘tumble’ and Hunt ‘negative talk’ warning” (link). Aren’t you glad? Never mind that Mr Shapps had already hinted last week that this would happen – it’s ‘News’ today.

Why we should ‘rejoice’ over the piffling sum of about £450 being taken off our bills which are now well into the £3,500, that’s the secret of our ‘betters’ who are incapable of ‘doing maff’. If, as reports outside the Westminster editorial chambers mentioned, the gas price has fallen below the sums at the start of the Ukrainian war, then surely there’s a digit missing in this fabulous sum of reduced energy price we are told to be happy about!

Now that I’ve got this angry grumble out of the way there’s another item which The Times suddenly seems very keen on: AI. There are two reports on their online front page this morning, one is the leading article at the time of writing. This is the headline: “AI ‘is clear and present danger to education’ – School leaders announce joint response to tech” (link, paywalled). I’ll point out in an aside that the accompanying photo is a bit off: how many male teachers are actually teaching in our schools? 

Anyway, the point of this article is that some head teachers have got together a conference about those AI dangers and wrote a letter to The Times. Here are some remarks which I found entertaining (all emphases are mine):

Head teachers’ fears go beyond AI’s potential to aid cheating, encompassing the impact on children’s mental and physical health and even the future of the teaching profession. Their letter says: “Schools are bewildered by the very fast rate of change in AI and seek secure guidance on the best way forward, but whose advice can we trust? We have no confidence that the large digital companies will be capable of regulating themselves in the interests of students, staff and schools and in the past the government has not shown itself capable or willing to do so.” (link, paywalled)

Yeah – bad government, bad Big Tech: we get it. Keep the question about whom to trust in mind and note well their worry about ‘the future of the teaching profession’. There’s more, e.g. when Mr Seldon who seems to be the chief letter writer warns that the greatest

“threat for teachers and pupils alike was “infantilisation”, because “AI will make life easy and strip away learning and teaching — unless we get ahead of it.” (link, paywalled). 

Never mind that during the lockdowns schools were shown to be in the digital stone age as all those teachers, cowering at home, were apparently incapable of providing digital content for their kids. Moreover, I well recall the howls about ‘deprivation’ when schools didn’t have enough tablets or laptops available in their schools. So on what computers  is this AI going to be run on? 

Here are two more quotes to keep for the record. Firstly Mr Seldon is saying that this is all developing ‘too rapidly and that

“there’s deep concern that the technology will be used for [children’s] learning, their stimulation, their care. The grounding of mental health is in relationships and human contact.” (link, paywalled).

Odd, isn’t it, that these arguments weren’t heard by teachers unions when kids were locked out of their schools during the covid years! And finally, after acknowledging that AI might ‘free’ teachers for more ‘interpersonal contacts’, 

“He added: “I’m worried young people are losing touch with reality, spending so much time interacting with augmented reality and virtual reality. As learning becomes more technological, their grasp of what is real and what is normal, what is true and untrue, will be progressively eroded. (link, paywalled)

Isn’t it already too late to worry about ‘young people losing touch with reality’ when they are already ‘interacting with virtual reality’ and have done so for years, thanks to being glued to their smartphones? Not that their parents are any better … 

Meanwhile, columnists in the Westminster press are trying hard to explain that AI isn’t ‘sentient’ while trying to show that ChatBots – which are AI – are quite useful when e.g. trying to dabble ‘in the markets’ (link, paywalled).

The one point not made by those ‘defenders’ of our teachers needing to actually teach, nor by the market dabblers nor by all the other clever tech reporters is that all these AI apps are having to be taught – by humans. There was a lovely little piece in the DT, reporting on ‘armies’ of people in Africa who are checking:

“[…] millions of pages of digitally transcribed text for accuracy, listen to voice assistant recordings and double check security camera alerts, correcting errors and perfecting responses. According to the Economist, one such Chinese data labelling company employed 300,000 people, carefully annotating a syllabus from which AI algorithms can learn. Data labelling jobs are fast becoming part of a new economy of machine learning work across the world. By some estimates it is already a multibillion dollar component of the AI industry.” (paywalled link)

I wonder if there are ‘data labelling jobs’ here in the UK: if not, why not? Are we Brits too lazy – or too ill-educated to do such jobs? I am not surprised that we hear nothing about who and which companies are working so hard on data labelling, making AI into the ‘solution’ for all sorts of problems, from BT’s customer service to ‘teaching’.

However, there was a brief report from the USA yesterday evening which should frighten all of us and which shows that there will be one problem not addressed by teachers and other ChatBot enthusiasts. According to this article, Mr Blinken, US Foreign Sec, announced in a speech that

“ […] the State Department has developed an AI-enabled online Ukraine Content Aggregator to collect verifiable Russian disinformation and then to share that with partners around the world.  […]  As a system that reflects the data on which it’s trained – including the biases embedded in that data – AI can, of course, amplify discrimination and enable abuses. It also runs the risk of strengthening autocratic governments, including by enabling them to exploit social media even more effectively to manipulate their people and sow division among and within their adversaries.” (paywalled link)

Why does that not scare the collective, ever so independent minds in Westminster? Is it because it’s ‘Teh USA – good, Russia – bad”? Which other agencies are or have already used ‘AI-enabled’ apps to ‘fact-check’ and disseminate info to us plebs? 

How can we trust anything we now read on the internet? How can we even trust government ‘announcements’ when we don’t know if they’re AI-generated, the input into this info having come from faceless people, perhaps sitting in Nairobi, employed by Chinese companies?

This should be the true reason for being wary of AI, not the wailing about children ‘losing touch with reality’, a touch most of them seem to have lost already thanks to their smartphone abuse, namely that we can no longer trust any info, any ‘fact-check’ because vested interests push their own agenda by using ‘AI-enabled news’. aimed at blanking out ‘fake’ news generated, of course, by ‘bad ChatBots’ ….

That’s all I have for today. So far, we’re assured, AI ain’t sentient, that it’s only about facts and not about feelings. Well, that’s good, innit, because we’ve seen over the last few years that ‘feelz’ and the alleged hurting of those feelz is what matters to our MSM. So AI cannot fake those … or can it? Have a good weekend.