I love this, and it has me fearing even more for the university students of today and the next decade. I also worry about k-12, there are innovative things happening at the fringes but the vast majority of public school systems are doing nothing to rethink education in a world with AI. And it’s the next generation that will suffer tremendously unless there are more dramatic shifts in the near term.
I'm a long time fan of Niall the historian but an outsider to academia. It is fascinating to see him interact with students. He obviously has a passion for education.
I'm 73 and these kids who don't read but expect AI to think for them will one day be running this country and the world which, unless we get a handle on this, means that AI will eventually replace these students after graduation and then what do we do with them? This "cloister" concept of which I approve would give students the requirement that THEY think instead of the machines.
Wow, your accomplishments truly humble me.. but in 1983, as I was a fresh BSEE I could see this happening as clear as day. Maybe it was all the science fiction I read when I was in a board Catholic Military school, or my fascination with all things novel and technical. Needless to say, I helped usher in digital cellular phones, broadband cable and twisted pair, WiFi, MPEG, JPEG, video acceleration, and OC-xx. And now I see the hype that is AI. Funny, I was hopeful that it could truly be useful, but it has been so corrected that sadly, I think it will never get past a 8th grade understanding of logic or thinking. Still, your article is spectacular and thank you for penning it.
Thinking and managing vast amounts of information- yes. But also INTEGRITY, which is in short supply and perhaps not well understood by young people. When anything is permissible to further an agenda or a narrative, integrity is lost.
Later than, better than, similar to The Robert E. cook Honors College at Indian University of Pennsylvania. A reunion for the 25th Anniversary of the first Cook Honors College in underway on June 20-22, 2025. The idea was the same, but UATX seems to have scrubbed the idea a bit with the help of an Oxford grad.
I have been a professor for 40 years and just retired. These LLMs are not the enemy of learning at all, BUT students need foundational skills and foundational knowledge to use this tool well. The difference is people actually care about the future, and LLMs do not. So those educated people who come to the LLM with a strong foundation of learning, thinking and analyzing as caring humans can create much more useful prompts and evaluate the responses from the LLMs much more effectively than novice students can. Student GPT papers are always benign if not fully tedious -- hard to prove any particular paper was cheating, but it is easy to know nearly every GPT paper submitted by a student will always be just another mediocre paper. And there is a reason why too: In a nutshell, Humans project a future; ChatGPT et al can at best predict a future. Humans live looking always into the future; LLMs exists in the perpetual present. Educated people simply dream up more interesting futures and have more complex worries and concerns about the future. We live there. We live always in a place that does not quite yet exist. We save money for example, students don't. That is all about the future.
Absolutely spectacular article.
Yep, if I was a college student again, then I would sign up for the cloister seven hours per day!
I love this, and it has me fearing even more for the university students of today and the next decade. I also worry about k-12, there are innovative things happening at the fringes but the vast majority of public school systems are doing nothing to rethink education in a world with AI. And it’s the next generation that will suffer tremendously unless there are more dramatic shifts in the near term.
I'm a long time fan of Niall the historian but an outsider to academia. It is fascinating to see him interact with students. He obviously has a passion for education.
The above essay by Niall Ferguson is not only technologically insightful but reads like a mesmerizing novel.
David-Ross Gerling, PhD
I'm 73 and these kids who don't read but expect AI to think for them will one day be running this country and the world which, unless we get a handle on this, means that AI will eventually replace these students after graduation and then what do we do with them? This "cloister" concept of which I approve would give students the requirement that THEY think instead of the machines.
Wow, your accomplishments truly humble me.. but in 1983, as I was a fresh BSEE I could see this happening as clear as day. Maybe it was all the science fiction I read when I was in a board Catholic Military school, or my fascination with all things novel and technical. Needless to say, I helped usher in digital cellular phones, broadband cable and twisted pair, WiFi, MPEG, JPEG, video acceleration, and OC-xx. And now I see the hype that is AI. Funny, I was hopeful that it could truly be useful, but it has been so corrected that sadly, I think it will never get past a 8th grade understanding of logic or thinking. Still, your article is spectacular and thank you for penning it.
You don't think all students should learn some history, some literature, some classics? Should it all be tech savvy?
Thinking and managing vast amounts of information- yes. But also INTEGRITY, which is in short supply and perhaps not well understood by young people. When anything is permissible to further an agenda or a narrative, integrity is lost.
Probably Alfred North Whitehead, as well as Oppenheimer.
Brilliant. If we just do this!
Later than, better than, similar to The Robert E. cook Honors College at Indian University of Pennsylvania. A reunion for the 25th Anniversary of the first Cook Honors College in underway on June 20-22, 2025. The idea was the same, but UATX seems to have scrubbed the idea a bit with the help of an Oxford grad.
I have been a professor for 40 years and just retired. These LLMs are not the enemy of learning at all, BUT students need foundational skills and foundational knowledge to use this tool well. The difference is people actually care about the future, and LLMs do not. So those educated people who come to the LLM with a strong foundation of learning, thinking and analyzing as caring humans can create much more useful prompts and evaluate the responses from the LLMs much more effectively than novice students can. Student GPT papers are always benign if not fully tedious -- hard to prove any particular paper was cheating, but it is easy to know nearly every GPT paper submitted by a student will always be just another mediocre paper. And there is a reason why too: In a nutshell, Humans project a future; ChatGPT et al can at best predict a future. Humans live looking always into the future; LLMs exists in the perpetual present. Educated people simply dream up more interesting futures and have more complex worries and concerns about the future. We live there. We live always in a place that does not quite yet exist. We save money for example, students don't. That is all about the future.