Engagement

Neither I nor anyone else can predict the future of AI. Let’s face it, even five years ago I had no idea how far the technology would become enmeshed in our everyday lives. I didn’t even realise how central it would become to the last couple of years of my full time teaching career. Predictions of where it will be in another five are, quite honestly, a mugs game. Rather than wasting our time trying to second guess the speed and extent of the AI revolution, we need to spend it deciding how to deal with it, and that means proper engagement.

Engagement with AI must start at school under the guidance of teachers who properly understand AI. As a profession we have been extremely bad in dealing with the technology because we have always seen it as a threat. It’s only a threat because the vast majority of education professionals refuse to deal with it in a balanced way. Children are very adept at picking up the mechanics of new technology, but appalling at understanding it and using it correctly and ethically. The blanket approach of dealing with AI as the devil incarnate creates a space for misuse and misinformation as children pick their way through a new landscape without so much as a map. So, I have set out the problem, what is the solution?

The solution starts in Primary School with engaged teachers understanding and teaching the basics of computer literacy. In an age appropriate way, as we should do with sex education, we set out what computers are and what they can do. We emphasise that they are nothing more than an advanced tool to speed up processes. They are not a technology to outsource our thinking to. For example, we could start with maths and say to the students, ‘Use your tablets to find out what 12 x 15 is.’ When the answer comes back as 180, we then ask the students to explain the process of multiplication that AI has just used. When they can’t do that we explain that if they rely on their tablets to do maths for them they will never be able to understand it. This is identical to the process my generation went through with calculators in the late 1970s and early 1980s when our maths teachers refused to allow us to use calculators in lessons or exams, and it is absolutely the correct approach.

We then move on to opinions and even facts which are famously malleable these days. The echo chamber isn’t just a feature of social media, it’s central to ChatGPT and similar AI tools which are built to be a sounding board and to broadly agree with you. How many times have you used an AI tool that has called you an idiot?! They are incredibly powerful and useful resources if you use them in the right way, but you can’t afford to accept what they give you back as facts or advice without checking it first. Just as you wouldn’t automatically trust a random stranger to give you a properly thought out and evidence based viewpoint so you shouldn’t automatically trust ChatGPT.

Children need to be taught and trained very early on, in controlled settings, to look for independent verification. It’s an easy process to adopt especially if it’s done early, and, if we don’t want the next generation of adults to be as clueless and gullible en masse as the current adult users who swallow any old rubbish because they saw it on Facebook, it has to start before their minds are outsourced to the machine.

As they go through school and university we increase the level of complexity but keep the same approach. Don’t automatically trust one source, look for others to confirm or contradict. If you see a newspaper article online, make sure you know what the paper’s broad political leaning is and who the owners are supporting financially or intellectually and why. If you see a peer reviewed article in a journal, ask yourself whether there are any questions left after reading and consider why those questions remain. Look at whether the evidence produced is compelling or at least persuasive and then use an AI tool like ChatPDF or Perplexity to gain an understanding of the evidence base and its position within the broad field of research. There are published peer reviewed articles arguing against anthropogenic climate change or the effectiveness of vaccination, but if you cast your net widely you realise that they are very much in the minority and very often reflect the views of a person or an organisation who employs that researcher. In order to reflect the views that they wish to, they might ignore certain evidence or put an unusual spin on it, meaning that they are not lying but laying out a viewpoint that is quite misleading when you look at it in detail.

So, in conclusion, I return to the point made in my introduction, namely that it is impossible to predict the future of AI even by the end of this decade. Whatever that future is, it has to be the result of engagement from everyone not just politicians or tech bros with profits to be made and influence to be bought. They have inflicted untold damage on society already with their selfish and immoral actions and the only way to stop things getting worse is to ensure that we have the tools in our own heads to question and disagree with these people. Engage with AI before it’s too late. Ignoring it was always a bad option and now it’s quite simply an existential one.