Ei-iE

Facts, truth and wisdom have been concerning me in recent times as we watch AI colonise the world. I fear we are losing our capacity to identify truth as we rush into a future where the algorithms present us ‘facts’ that may or may not be true.

Once we could reliably turn to the academic literature, the published works of academics, to find the current knowledge on a topic and to trace the process of discovery over time. A work would have been written by scholars, backed by substantial research, and reviewed by peers before being published. Regardless of the quality of the work, it had been through an identifiable quality control process prior to publication. Our governments and our communities rely on this information to make critical decisions.

But now, in the age of the algorithm we are seeing publications where the work is not guaranteed to be generated by humans. AI is being used to ‘correct’ writing, to compile references, to draw conclusions from data and in some cases to generate data to support an idea. Apart from a commitment that the ‘person’ presenting the work for publication has made about the content, there is no guarantee anymore that what we read is presenting us facts based on legitimate data.

In Australia we have had union members tell us that their published work has been stripped and the data used in other publications, that they have been cited as authors of work that they didn’t do, and as authors of work that may not even exist. For academics this is a serious challenge to what is the currency of their career, their publication record and the impact it makes through citations. Not only that, the inherent bias in AI also shows itself here when obviously female names of authors are changed to male versions. In addition to this, some publishers have not only sold the right to mine their publications to the big tech companies but have also demanded that authors of academic books grant them the right to ‘update’ content at any time. This is not just a problem in Australia; this is a global phenomenon.

And it leads us to the question: if we can’t rely on the academic literature to provide facts to determine the truth, where will we find it? One of our local solutions is to call on our government to establish a repository of all Australian academic works, which can be a reliable source of information. Think of it as akin to the seed bank established to preserve the world’s plants, only in this case to preserve research findings, the germs of discovery. This also needs to go hand in hand with the development of sovereign AI capacity to preserve Australian culture, traditions and language, especially that of the First Nations people of Australia, the Aboriginal and Torres Strait Islander people.

The Large Language Models driving generative AI reflect their origin, the values of their developers and the balance of materials available on the internet. As everything is homogenised through GenAI, we risk losing our global diversity. That our students at all levels are relying on AI and seeing the world through that lens with all its bias is deeply concerning. At the university level where we work, we are seeing the impacts with students losing either the capacity or the will; to reflect on, verify, and critique the outputs that GenAI is producing for them. We are also seeing a decline in creative and innovative thinking in student responses to assessments.

Over several decades our higher education system has been subject to mass expansion and the commodification of education. We are chronically under-funded, and under-staffed (with first year class sizes of more than 1000 students not uncommon), and our workplaces are not mentally healthy. Introducing AI into this environment without proper consultation is and was a recipe for disaster, that we are seeing the beginnings of now.

It is beyond comprehension that the current leaders of our universities, the temporary custodians of important institutions with a deep history, are prepared to give away the competitive advantage that they have over Big Tech. The core of a university education has always been people; the human relationships, the lecturers and tutors who mentor and support, the fellow students with whom you passionately debate ideas, and the opening of your mind to new ideas and ways of assessing those ideas critically. By replacing staff with chat-bots, by moving all content online so there is no imperative to be physically present, by encouraging the use of AI tools that we know reduce critical thinking skills, and by cracking down on free speech and protest, the universities have been hollowed out and the experience, once so rich for students, is slowly being impoverished. What is the advantage to being educated on an empty campus when you can get a degree online from BigTech?

Apart from lobbying our government to take steps to protect Australian universities, the current state of research, and the knowledge that will be generated in the future by our students, we are tackling this through collective endeavour. We are currently pursuing AI clauses in our 50+ collective agreements, that will require: extensive consultation prior to introduction of AI, an identifiable human decision maker when algorithms are used to assess workers, protection of the moral rights of academics, and complete transparency about the when, where, and how AI us being used by our employer. If AI must be used, it must be used to augment our work not to replace us. This last point is particularly critical for education support personnel who are at the front-line of job losses that have already begun.

We must do everything we can to ensure we are training a new generation of independent critical thinkers through meaningful human interactions, to protect the global knowledge base now and into the future, to, and ultimately, to protect facts and support truth.