I just started Law school at a university with a prestigious history. AI use is rampant and very predictably is delivering a safety net early in the program. My issue with this is that it is shifting the locus of attention in education further toward assessment results and further away from individual learning. Law and Med school are particularly vulnerable to this due to the impact exit grades have on early career opportunities.
Particularly sympathetic to people on this forum, my concern with AI use in higher education is the long-term impacts of outsourced cognition. The issues for the universities themselves are a bit trifling and higher ed had structural issues with performance and technology far before AI. The risk to the individual in terms of advanced cognitive disorder onset remains to be seen but could be a consequence (reduced neural network density and recall from cognitive outsourcing leads to early onset cognitive decline, see most dementia research and the MIT study on AI use in essay writing (brainonllm . com)).
The actual risk to me is inculcation into a profession that has decision impact in society is being made dependent on proprietary technology which increasingly has political implications.
I’m much more worried that in the context of a scarce computing event (people loose access to high performance computing and network technology – not too out there with resource and supply chain shocks) people will be incapable of maintaining a profession’s minimum operational level for society to function and move through the event. Continuity of professional experience is also a deep concern in this regard. If junior position professionals are outsourcing tasks to AI instead of learning from senior professionals, the tacit knowledge loss could be quite dire. Knowledge management issues of tacit and explicit knowledge value gaps is an extensive area of study in management sciences and should be talked about more. AI can only deliver explicit knowledge guidance, if not underpinned by tacit experience, the value of that explicit knowledge deliver is extremely limited.
I am more concerned that people going into higher education and the people supposedly advising them do not understand the importance of training and educating for worst case scenarios. On a philosophical level, in my little world, training to be a jurist (lawyer) means that if I was dropped in a settlement situation with a group of people with few resources, I would know and understand enough principles of law and procedure to establish a system of civics fit to that group’s needs and establish some level of order within a reasonable time frame of a year or two.
If someone relies on cognitive outsourcing to extract, summarize, and synthesize throughout their studies, they will be reliant on that tool to perform what they were trained in. Try having a carpenter trained on power tools build a table with hand tools, it won’t be a smooth process. We should educate for a profession which is then performed in an economy, not educate for an economy that performs different professions.
There is a difference between being educated with AI and a calculator, scale. Calculators are readily deliverable for minimal production and maintenance costs, even an abacus and a slide rule can substitute for all but modeling calculations if people know how to use them. AI is extremely expensive to develop, produce, build out the infrastructure for, distribute, maintain, and learn. That is my worry. Calculators didn’t destroy numeracy but they also are tools which adapted the same role from existing technology, and they did decidedly impact arithmetic fluency for the individual. AI is a full systemic sea change of knowledge in society and requires extremely expensive upkeep, both resource costs and for labor hours. The only outcome to me for continued AI adoption is technological reliance which breeds political allegiance out of necessity to access tools for basic processes.
I think the proof will be in the pudding in 15-20 years of what AI and extensive technological adoption does to human cognition and social functions. Until technology is proven safe through extensive longitudinal studies, I suggest viewing and advocating against it as a workplace safety hazard, it is literally outsourcing your professional experience and educational opportunity for a productivity enhancement that benefits someone else, you should be compensated for that loss.
We have come to a place in society that we have deemed it necessary to have some level of testing for new drugs to ensure a modicum of safety for use and consumption when the benefits outweigh the consequences. Technology should be no different.
You can also take high doses of amphetamine for short term task specific performance benefits but it has long-term consequences on adrenal health, memory, and sleep architecture. I am happy to take the performance hit and actually live my life and enjoy my education as the wonderful process of learning and becoming that it is. Relying on AI in higher education is like selling a kidney to make a down payment on a house, you might get a house sooner but you’re probably not living in it as long. That’s my soapbox. Choose a life. Don’t choose a life to outsource it. I would rather be a learned beggar than a vapid cushioned shill simpering to the repetition of task and prompt slave to some screen.
Honor to Mnemosyne, mother of muses. May the springs of the pool of memory run eternal.
