AI as a writing teacher?

I will participate in a debate about AI in the teaching of writing. This is my opening remarks.

Link to the Norwegian version.

Whether we should use AI in teaching writing almost always boils down to assessing small advantages against small disadvantages. For example, we can hear:

"The advantages of language models are that they can help students with language cleaning, spelling, writing frames, and give feedback for further processing."

The disadvantages can be described as follows:

"Students cheat and they learn nothing. I have become an AI policeman.”

But this pro et contra exercise is rooted in a narrow understanding of what teaching writing is about.

A meaningful discussion of the advantages and disadvantages of AI in teaching writing must look at long-term consequences, not just at the processes that occur at the moment of writing.

Learning to write is a lifetime project. As a writer and a teacher, I still think it’s the most difficult thing to do.

If we only consider short-term perspectives, our assessments do not reflect what is at stake - and it is not whether a student can place a comma without the help of ChatGPT.

It is, to put it dramatically, the death of an entire culture we are talking about.

What, if not short-sighted digitalization, is the reason why governments are suddenly allocating money for school books in many countries right now? Did no one think how important it is to be able to read in a democracy?

Short-term perspectives almost always bring with them unwanted and unseen consequences. And I'm sure we will also get that because of AI in schools when elementary schools ask children to ask their AI friends for help.

We have to think in much longer perspectives. But due to the fact that we live in an incredibly accelerated time, it will do to look a few decades ahead.

What do we see then?

Or to put it another way: Is the fundamental dilemma really whether AI tools give students a marginally better writing education?

If we answer yes, I think we are professional idiots trapped in a dangerous silo thinking.

The real dilemma for the school, and thus the teaching of writing, is how to prepare students to live in a future where AI is everywhere. What should students learn? How should they reflect on this? What other risks interfere with AI? How will they be able to make informed decisions about what is best for them to do?

Before I get to the point, I need to say a few words about what I think writing is.

Or put another way: What is being practiced when we practice writing?

On a superficial level, writing training is training to communicate. We learn to create a surface that conveys meaning.

But on a deeper level, it's training in understanding what you want to communicate -training in thinking clearly.

Writing training is therefore a painstaking and slow process where we who write gradually discover who we are, what our identity is, and who we can be. For young people like our students, writing training is incredibly important - quite simply a radical opportunity.

The formative aspect of writing training is on the wane in a public that has developed a more and more instrumental view of teaching and writing.

The saying goes something like this: "School is necessary training for future adults to be able to perform a job." Or: "Good writing skills are necessary to the extent that they will be needed in the professions of the future - and that is unlikely since we have AI models that already write superbly!"

But is writing primarily an appropriate medium for transmitting information - a medium we switch from until there is something better, which AI will surely invent for us?

No. Writing is the dialogue we have with ourselves to become who we are.

Writing is a transition between daring to formulate an unthought thought, and then rejecting or refining it into something you can stand for.

So this is my starting point. Writing (and reading, of course), is essential for being able to think independently, which in turn is essential to think truly critically, which in turn will be essential in preparing for the future.

Okay. What characterizes our future prospects, if we think a little further ahead than our students’ next writing assignment?

I think the prospects are bleak, and the debate about the use of AI in writing education reveals that school pupils hardly get a shred of insight into what awaits them after school.

Before we ask ourselves about the benefits of AI in writing training, we must think through more fundamental questions, and ensure that critical thinking does not lose ground.

I will discuss two aspects to show where the shoe presses: energy consumption and surveillance.

The climate crisis is raging towards us at enormous speed. While the IPCC believes we must halve emissions by 2030 and be at net zero by 2050, emissions are increasing.

At the same time, AI is an industry into which investments are being poured, and with which the fossil fuel companies are also integrating. AI lobbyists nurture false hope that AI will help us with EVERYTHING.

Yes. AI will also solve the climate problems!

But how likely is that? This year there have been more and more warnings about how much electricity data centers need. In Norway, we have a debate about a TikTok center, which alone is predicted to consume 1 percent of all electricity in Norway.

But AI is supposed to help us make energy consumption more efficient, say the lobbyists.

Yes. However, all experience indicates that the efficiency gain will be used to increase consumption and scale up even more.

The climate crisis will probably get worse with AI.

From an energy perspective, a huge and growing AI industry is exactly what we do NOT need.

But what do students know about that?

Surveillance is another scary area. Our students are already totally supervised to the extent that many are numb and indifferent. But does that mean we should accept the situation?

Many warn that we are facing completely new forms of society when surveillance with facial recognition, room modeling via wifi signals, and the traces we knowingly and willfully leave behind around the clock are integrated into total AI surveillance systems.

Daniel Schmactenberger believes that it is still possible to change the course of the future, but that we will soon be at a point where popular rebellion is impossible.

Yuval Noah Harari claims in his new book, Nexus, that the only relevant comparison with AI is the monotheistic religions from old age. Only with one important difference: the AI ​​religion will be much more powerful than any religions we've seen before.

The AI ​​religion, quite unlike the Koran, the Torah, and the Bible, will be able to interpret itself and create new societies, where people, totally monitored even in their sleep, do not make a single independent choice.

I could have gone into many more terrifying scenarios that AI puts forward, but I'll settle for these two.

The question then becomes: How should teachers and their students approach this?

Well, the students must at least have the opportunity to have a critical attitude to the future prospects that AI outlines - we cannot let ourselves be blinded by marketing.

This means that if teachers are to take a position on the use of AI in writing training, then we should be willing to think much more radically than the debate has shown so far - as the change of words is still within the framework of the school as we know it.

For AI in schools to make sense, we must rethink writing's potential to develop competence that simultaneously cultivates long-term perspectives - and thus form critical citizens who can help us protect what we want to take care of in the future.

For example, should we drop grades, since it's a pointless exercise with AI?

Without grades, you can spend zero time on compulsory writing tasks that will only provide a basis for grades, and work much more on the general educational potential of writing.

Then it somehow becomes irrelevant whether ChatGPT can help the student put commas in the right place.

Prepare for debate!