Is AI just a tool?

Pondering the TLWAI Europe panel session

I recently spent an excellent two days at the first Teaching and Learning with AI in Europe conference, organised by Breda University of Applied Sciences and the University of Central Florida, so first I want to say a big thank you to everyone who helped organize the event, those who presented and also to all those who talked with me, thank you for the interactions and thinking. There were a few things that were raised in the panel session that we did not have time to really explore and that I have been pondering since the event and so I have decided to write that up as a way of thinking these things through.

Why I do not agree that AI is “just a tool”

In the response to the very first question to the panel it was said that AI was ‘just a tool’ and that position was shared pretty well by the rest of the panel except me, so I would like in this post to explain more fully than time allowed, why I do not share this view but instead think that AI is a technology in which we all participate.

AI is a technology in which we all participate.

In my initial response I mentioned how AI is shaping society and how we interact with it and each other, how the tendency to anthropomorphize means we often, for example, thank it for help, where we would not do that with a hammer and I am aware that this comment did get a few laughs and also nods of agreement. I think I mentioned briefly Heidegger. But there was of course not time to really get into fully explaining any of this in detail.

The hammer

So, to start with the hammer, and anthropomorphism. AI is designed to encourage a relationship with the humans who use it, in some cases it is named, and the entire design of our interaction invites us to see it as a person. This is the basis of the Turing test after all; can we reliably tell whether we are talking to a human or a machine? Current AI design is to make it harder for us to be able to tell. This is a commercial choice to shape how we think and engage with it. Traditional tools are not designed in this way. Further, the talk from AI firms of consciousness and needing to protect them from possibly harmful interactions with humans reinforces that position.

To dig deeper into my use of hammer and mention of Heidegger, he was trying to show that technologies were not just a means to an end but in fact reveal what things are and what they are for. They enframe things and his concern was that this enframing was showing the world to be a ‘standing reserve’ – in other words, things to be used. Calling AI just a tool is thus just the sort of enframing that Heidegger was warning us about.

Tools and technologies

I want to add that I think that there is a difference between tools and technologies. Here I am drawing on ideas and work by Brian Arthur and Jon Dron. A tool is something we use, technology is an orchestration of phenomena to achieve some purpose. We humans often are a part of that orchestration. So, I use my hammer as part of the orchestrations of phenomena to assemble an Ikea wardrobe, the hammer is part of a technology that includes me. Dron calls the human part of this orchestration, where we flexibly improvise and judge as we go, soft technique. Hard technique, by contrast, is the mechanical, routine, the scripted. Machines have been able to do hard technique for a long time, but AI is arguably the first widely available technology that can do soft technique and that is important. It writes, it evaluates and it converses, a screwdriver, a hammer and a calculator never did that.

Not neutral

I mentioned that AI does things to us. The panel did not disagree with that point, but argued that it does these things as a tool. I disagree because using the word tool frames AI as neutral, and it is certainly not. Here I am drawing on work by Neil Selwyn among others. Tools do things, technologies do things and do things to us. These technologies are being embedded into other technologies so that they do not only shape us but start to shape our other technologies. If we frame AI as merely a tool then we are hiding this and framing some sort of neutrality that is not real because it carries, embedded in its very design, the beliefs and values of the industry and people who built it. Those cannot be removed.

Tools do things, technologies do things and do things to us.

Those assumptions about the use of data, about efficiency, about infrastructure, about what counts as knowledge, or what counts as valuable, are all there in the technology. This makes it political and social in ways that are often obscured when we describe AI merely as a tool.

My aim here has really not been to try to win the panel or to make a point, rather the act of writing this is my own thinking in action as I try to reach for what it was I felt important to say but had neither the time nor the skill on the day. This is why this is entirely my own writing, AI has not been used. Errors in facts, in grammar and spelling and in thinking are all my own.

References

Arthur, W. B. (2009). The nature of technology: What it is and how it evolves. Free Press.

Dron, J. (2023). How education works: Teaching, technology, and technique. AU Press.

Heidegger, M. (1977). The question concerning technology, and other essays (W. Lovitt, Trans.). Harper & Row. (Original work published 1954)

Selwyn, N. (2016). Is technology good for education? Polity Press.

Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460.

^{Photo by}^{Clarisse Croset}^on^Unsplash

Writing Something at Last

This is more of a “link blog” than anything reflective and in fact it is really just a lightly edited version of a page that was used in a course that I was teaching in Emerging Trends in Education. However, perhaps it does serve as a way to reopen this blog and my writing. Perhaps there are some links here that might be useful.

Quote of the week.

“The real danger is not that computers will begin to think like men but that men will begin to think like computers” (Sydney Harris – Journalist)

Jon Dron quickly mentioned the quote during his recent webinar and so brought it to my attention.

How Good are you at Identifying Fake GAI Images?

Try this to find out. Might be a useful resource to use with students as well.

No Need to Worry – Everything is Going to be Wonderful

So says Sam Altman, Head of Open AI in his recent blog post. Why? Because “Deep Learning Worked”

AI in Early Years and School Settings

Much of the literature shared on AI in education has a higher education focus. Happy to see this from the Teacher Learning Network (TLN) in Australia which provides “high quality professional development for staff in early childhood, primary and secondary school settings.”

Link Here to home page and PDF of the Issue here.

UNESCO AI Competencies for Teachers

A new report. It does include a short summary for the TL/DR.

Funding for AI in Education

Google.org announces new AI funding for students and educators. I had personal experience of this sort of thing with Microsoft. A Faustian deal or a useful source of money and PD, both of those or neither? You decide.

Call for Papers

Metaphors of AI in Higher Education – Discourses, Histories and Practices: Call for Papers. If you have ideas and want to talk them through, happy to do that.

Contrarian Views?

There have been an increasing number of calls for the development and enactment of “AI Literacy” and this almost always includes a focus on the ethics of AI. Punya Mishra questions that assumption.

Bloom’s Two Sigma from the 1980’s is pretty well embedded as a given and often cited to support arguments in education. Just how valid and reliable is the research and has it been repeatable? This article takes a look.

Both of these are shared in the spirit of being “critical” and questioning assumptions.

Where Good Ideas Come From?

This is not so new, but rather it is only recently that I made a connection. One of the Substacks that I read is by Steven Johnson. He is also one of the main people working at Google and behind their NotebookLM project.

What I had not connected is that this is the same Steven Johnson who wrote “How We Got to Now” and also “Where Good Ideas Come From“. Both of these I had read some years ago but I just had not made the connection that it was the same guy.

This led me to revisit “Where Good Ideas Come From” because I think there are some interesting insights in that book, some of which have also been built upon by others. For example, he writes a whole chapter on the The Adjacent Possible and another on Liquid Networks. Some of you will no doubt recall that Jon Dron also writes and speaks about the Adjacent Possible and the role of Path Dependencies, in his papers, books and presentations. I also think that my own thinking about a Liquid Curriculum and also Networked Knowledge is influenced by Johnson’s writing in this book, although I was not seeing the source of my own ideas.

If you do not have time for the book or just prefer video then here are some resources:

This 4 minute video focusses on The Slow Hunch and the importance of Connectivity.

This is 17 minute Ted Talk Video, provides greater detail.

Although these are more than a decade old, I think they might be useful. To bring this right up to date, this Sept 2024 article from Verge picks up the story with NotebookLM and how he envisions a Chatbot as a teacher! Now no matter what you or I think of that and your natural objections, we have to engage with these narratives.

Punya Mishra (of TPACK fame and many articles and books) – does not agree and explains Why ChatGPT Isn’t Your Next Teacher.

Category: AI

Just a tool?