I want to challenge one of the underlying assumptions contained in this post, an assumption that has been repeated as fact without critical consideration – the assumption that the current adoption of LLM Gen AI is only the beginning of a larger trend of widespread adoption. This is absolutely not a foregone conclusion. The money being lost on OpenAI is absolutely unprecedented. Open AI lost $5 billion in 2024 and loses money on even its paid subscription services. This technology is only offered to the public through absolutely massive amounts of speculative investment. Offering a cool tool to the public in hopes that it will someday become essential will only satisfy investors for so long. They will want to see user data first and then profits.
Regarding user data, according to OpenAI's own propaganda piece, Building an AI Ready Workforce, “More than any other use case, more than any other kind of user, college-aged young adults in the US are embracing ChatGPT…” No wonder OpenAI is doubling down on providing AI to students. Adoption by students is driving their growth. By contrast, despite business leaders’ excitement about AI generally, the actual workers tasked with using it have mostly not found much help from a general information device that is frequently inaccurate and lacks privacy protections. Few jobs are general. Most work is particular. Even in areas like customer service, there have not been any of the anticipated business disruptions. So, what industry has been disrupted? There are a number of business indicators that now suggest that AI has disrupted what I would consider to be the cheating industry. Companies like Chegg and Quizlet appear to be losing market share as ChatGPT gains. User trends revealed through Google Explore show ChatGPT use patterns over the school year with almost identical peaks and valleys as Quizlet and Chegg. More than any other use, ChatGPT appears to be used as a highly effective cheating device for students.
From a business point of view, what we have are AI companies building user numbers by tempting students away from the work of studying for grades, then using these growth statistics to convince businesses that they need to keep up with the users of the future (students), then trying to convince our institutions that we need to make our students AI-ready for future jobs.
In the meantime, those students who use the tool often see their learning compromised as a result. To me, this is a situation like Big Tobacco, where the companies are trying to build business by creating unhealthy dependencies.
I agree that we need to engage with this problem rather than sit on the sidelines, but I don’t agree that the future of AI is either clear or inevitable. We should advocate to hold these companies responsible for the harm they are imposing on our education system.
I think you certainly have a point about building dependency and I think most of us have seen first hand how students use this technology with no safeguards or ethical standards. In part, that's what these conversations need to spark.
I'm not entirely sure what's happened these past few months, but OpenAI supposedly had explosive growth that's simply hard to believe. ChatGPT users went from 250 million weekly in October to over 400 million weekly users by the end of February. Their paid for service also recently saw a 30% increase in revenue last quarter alone. I'd take this all with some deep skepticism as these are internally reported numbers, but they are making their rounds in the press and certainly driving interest.
What you can see by looking at search results in Google Trends is consistent with a student cheating theory. The current uptick may reflect testing and papers that were due before spring break. Also, if you look at the trends for Quizlet and Chegg, you will see their usage statistics falling, reflecting the theory that ChatGPT is growing partly at the expense of these other (cheating) companies. If you continue to watch the trends, we should expect that about mid-May, Google searches for ChatGPT in the US will start to fall dramatically as summer break starts.
For more on the business fundamentals, Ed Zitron provides pretty compelling evidence of a current AI bubble. For now, I will point out that there is a big difference between (unverified claims of) "revenue" and "profits".
When people make comparisons to previous technological changes, it's common to only consider those technologies that survived and are still in use. How do we know that this is not another Concorde Airliner? Another good comparison that is a bit more recent would be the 3D printer. Cool technology with a lot of initial hype but not a lot of business uses for it. And then there are the really scary comparisons, like Theranos.
I agree that educators can't sit this one out. In fact, hiding our head in the sand is irresponsible when it comes to preparing the future that students face. I am suggesting that there is a very different way to deal with the issue -- I'm suggesting adopting a "Proficiency" orientation. Here's a link to the ]Substack I posted earlier this week. I'd love to have your feedback on the concept I'm not wedded to all the details, but I think the concept has merit as an alternative to the yes/no discussions that many seem to be having. https://twelchky.substack.com/p/ai-in-schools-a-call-for-a-new-kind?r=6jqjj
I think Marc Watkins’s lead-in useful. I’m thankful for the flag about the new souped-up cheating machine trial period marketing to students.
The guest essay that followed, I found honestly rather naive, meandering, and verbose as cover for too much vacuous hand waving. Then I got to the AI disclosure at the end. Of course, I might’ve suspected that. But I wasn’t looking for it in a Watkins blog post.
As an AI Physicist, may I respectfully suggest the free downloads at https://phet.colorado.edu/. 1.7 billion students have downloaded the course material based on Active Learning! I met these teachers as we developed emergent learning methods for the most advanced small team honors students.
A large part of the 'white male based' orator approach to teaching is the long winded lecture. Dr. Wieman, Nobel Prize winner, provrd a radical alteration to the entire university structure of 50 minute lectures. Here, lectures are 10 minutes long. Studues show math retention >85% after a two year gap in math courses. Of course, extrovert orators with tenure are very dismayed. The revolution has started...On the ai front, please review the most advanced robot scientist in the world. 18 minutes Ted Talk https://youtu.be/4ht3m6p8iZ0?si=rsH0p5JKAvidbJ7o
I want to challenge one of the underlying assumptions contained in this post, an assumption that has been repeated as fact without critical consideration – the assumption that the current adoption of LLM Gen AI is only the beginning of a larger trend of widespread adoption. This is absolutely not a foregone conclusion. The money being lost on OpenAI is absolutely unprecedented. Open AI lost $5 billion in 2024 and loses money on even its paid subscription services. This technology is only offered to the public through absolutely massive amounts of speculative investment. Offering a cool tool to the public in hopes that it will someday become essential will only satisfy investors for so long. They will want to see user data first and then profits.
Regarding user data, according to OpenAI's own propaganda piece, Building an AI Ready Workforce, “More than any other use case, more than any other kind of user, college-aged young adults in the US are embracing ChatGPT…” No wonder OpenAI is doubling down on providing AI to students. Adoption by students is driving their growth. By contrast, despite business leaders’ excitement about AI generally, the actual workers tasked with using it have mostly not found much help from a general information device that is frequently inaccurate and lacks privacy protections. Few jobs are general. Most work is particular. Even in areas like customer service, there have not been any of the anticipated business disruptions. So, what industry has been disrupted? There are a number of business indicators that now suggest that AI has disrupted what I would consider to be the cheating industry. Companies like Chegg and Quizlet appear to be losing market share as ChatGPT gains. User trends revealed through Google Explore show ChatGPT use patterns over the school year with almost identical peaks and valleys as Quizlet and Chegg. More than any other use, ChatGPT appears to be used as a highly effective cheating device for students.
From a business point of view, what we have are AI companies building user numbers by tempting students away from the work of studying for grades, then using these growth statistics to convince businesses that they need to keep up with the users of the future (students), then trying to convince our institutions that we need to make our students AI-ready for future jobs.
In the meantime, those students who use the tool often see their learning compromised as a result. To me, this is a situation like Big Tobacco, where the companies are trying to build business by creating unhealthy dependencies.
I agree that we need to engage with this problem rather than sit on the sidelines, but I don’t agree that the future of AI is either clear or inevitable. We should advocate to hold these companies responsible for the harm they are imposing on our education system.
I think you certainly have a point about building dependency and I think most of us have seen first hand how students use this technology with no safeguards or ethical standards. In part, that's what these conversations need to spark.
I'm not entirely sure what's happened these past few months, but OpenAI supposedly had explosive growth that's simply hard to believe. ChatGPT users went from 250 million weekly in October to over 400 million weekly users by the end of February. Their paid for service also recently saw a 30% increase in revenue last quarter alone. I'd take this all with some deep skepticism as these are internally reported numbers, but they are making their rounds in the press and certainly driving interest.
Sources for this are below.
30% revenue growth: https://finance.yahoo.com/news/openais-revenue-set-triple-jump-173205114.html
Number of users: https://www.wsj.com/tech/ai/chatgpt-ai-cheating-students-97075d3c?st=ADYgGi&reflink=mobilewebshare_permalink
Correction: I wrote "Google Explore" but meant "Google Trends". https://trends.google.com/trends/explore?date=2024-02-05%202025-04-05&geo=US&q=chatgpt&hl=en
What you can see by looking at search results in Google Trends is consistent with a student cheating theory. The current uptick may reflect testing and papers that were due before spring break. Also, if you look at the trends for Quizlet and Chegg, you will see their usage statistics falling, reflecting the theory that ChatGPT is growing partly at the expense of these other (cheating) companies. If you continue to watch the trends, we should expect that about mid-May, Google searches for ChatGPT in the US will start to fall dramatically as summer break starts.
For more on the business fundamentals, Ed Zitron provides pretty compelling evidence of a current AI bubble. For now, I will point out that there is a big difference between (unverified claims of) "revenue" and "profits".
When people make comparisons to previous technological changes, it's common to only consider those technologies that survived and are still in use. How do we know that this is not another Concorde Airliner? Another good comparison that is a bit more recent would be the 3D printer. Cool technology with a lot of initial hype but not a lot of business uses for it. And then there are the really scary comparisons, like Theranos.
I agree that educators can't sit this one out. In fact, hiding our head in the sand is irresponsible when it comes to preparing the future that students face. I am suggesting that there is a very different way to deal with the issue -- I'm suggesting adopting a "Proficiency" orientation. Here's a link to the ]Substack I posted earlier this week. I'd love to have your feedback on the concept I'm not wedded to all the details, but I think the concept has merit as an alternative to the yes/no discussions that many seem to be having. https://twelchky.substack.com/p/ai-in-schools-a-call-for-a-new-kind?r=6jqjj
Bravo. Thank you for this piece. Now, let's see who heeds your advice.
I think Marc Watkins’s lead-in useful. I’m thankful for the flag about the new souped-up cheating machine trial period marketing to students.
The guest essay that followed, I found honestly rather naive, meandering, and verbose as cover for too much vacuous hand waving. Then I got to the AI disclosure at the end. Of course, I might’ve suspected that. But I wasn’t looking for it in a Watkins blog post.
As an AI Physicist, may I respectfully suggest the free downloads at https://phet.colorado.edu/. 1.7 billion students have downloaded the course material based on Active Learning! I met these teachers as we developed emergent learning methods for the most advanced small team honors students.
A large part of the 'white male based' orator approach to teaching is the long winded lecture. Dr. Wieman, Nobel Prize winner, provrd a radical alteration to the entire university structure of 50 minute lectures. Here, lectures are 10 minutes long. Studues show math retention >85% after a two year gap in math courses. Of course, extrovert orators with tenure are very dismayed. The revolution has started...On the ai front, please review the most advanced robot scientist in the world. 18 minutes Ted Talk https://youtu.be/4ht3m6p8iZ0?si=rsH0p5JKAvidbJ7o