YOU WON’T GET THE RIGHT ANSWERS…�IF YOU DON’T ASK THE RIGHT QUESTIONS

WHY CHATGPT IS COMING TO DESTROY HIGHER EDUCATION

MY GOAL IS TO NOT START RANTING UNTIL THE 5 MINUTE MARK

ABOUT THAT TITLE…

• If we only ask “how can we use this in our classrooms?” or “What widgets can I put in my canvas page” we will get one set of answers.
• If we ask “what is the end use of this technology” or “does use of ChatGPT lower critical thinking skills” or “will ChatGPT be used to replace all faculty” we get different answers.
• If we ASSUME “this technology is inevitable” we never ask “IS this technology inevitable” or “what about this technology has everyone so worried about it ending all life on earth?” (Yes, that is a real fear that those WORKING on the projects have)
• Three facts (“inconvenient truths”) and two implications

• Jumping from the bottom box to the top box is what makes “AI” such a big deal. The lower-ordered tasks aren’t “intelligent,” and the higher-level ones are important precisely because they can do things the user didn’t expect.
• So what happens if you ask an AI to “hack into the social security data base?” or “Make a biological weapon that only kills Latinx people?
• Before this was a “brute force” task – but now the AI can come up with a bunch of attack plans and test them out at super-speed trying an infinite number of discrete solutions.

SO WHAT IS A.I., EXACTLY?

COGNITION

THAT’S WHAT IT IS – WHAT DOES RESEARCH SHOW?��THREE INESCAPABLE FACTS

• The best available research shows that LLM use by students LOWERS critical thinking ability
• All available evidence shows LLMs are being developed in secrecy, the risks are being hidden and under-estimated in public debates, and the developers ARE NOT open to any outside entity encouraging “responsible” development
• AI is being deployed to CUT the workforce and INCREASE the workload of remaining workers. (This is not a route to our students getting jobs.)

INESCAPABLE FACT #1: THE BEST EVIDENCE SHOWS LLM USE LOWERS CRITICAL THINKING SKILLS

• Zhai, Zibowo, and Li (June 18, 2024) report that “By using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, our systematic review evaluated a body of literature addressing the contributing factors and effects of such over-reliance within educational and research contexts” and discovered “Despite the undeniable advantages of AI dialogue systems in streamlining research processes and enhancing academic efficiency, our analysis reveals a concerning trend: the potential erosion of critical cognitive skills due to ethical challenges such as misinformation, algorithmic biases, plagiarism, privacy breaches, and transparency issues.”
• Note it is the only comprehensive review of all data. This is better than single articles.
• It might do some other stuff, but IT LOWERS CRITICAL THINKING SKILLS.
• This is the logical extension of the computer taking over higher-order cognitive skills (like thesis, support, and conclusions) instead of lower-order skills (like spelling and grammar).

INESCAPABLE FACT #2 : IT’S SUPER-DANGEROUS, IT’S BEING DEVELOPED IN SECRET AND WITH NO OVERSIGHT (SLIDE #1)

• There is a group of AI developers – NOT Luddites, but the people WORKING on the tech – who have become whistleblowers. Their conclusions it is DANGEROUS.
• Whistleblowers developing the systems have repeatedly warned of algorithmic bias that increases existing inequalities.  In a document since taken down by the Trump administration, the US government itself has warned that ““irresponsible use could exacerbate societal harms such as fraud, discrimination, bias, and disinformation; displace and disempower workers; stifle competition; and pose risks to national security” (see open letter, footnote 4).
• Other sources saying the same thing are the UK government and the Bletchley Declaration of 29 countries, 3 tech companies themselves (OpenAI, Google DeepMind, Anthropic), and AI experts (250 signatories on Facct, 1,000 signatories on CAIS).
• DeepSeek will produce no criticism of the Chinese government human rights record and will instead reproduce communist state propaganda. If you don’t believe me, go try for yourself.

INESCAPABLE FACT #2: IT’S SUPER-DANGEROUS, IT’S BEING DEVELOPED IN SECRET AND WITH NO OVERSIGHT (SLIDE #2)

• Whistleblowers developing the systems have repeatedly warned of secrecy and the active suppression of safety concerns, hiding significant additional dangers from public view and skewing public debate in a way that seriously understates risks.  As one recent open letter, signed by 16 current and former AI employees warn, “AI companies possess substantial non-public information…We do not think they can all be relied upon to share it voluntarily…current and former employees are among the few people who can hold them accountable to the public. Yet broad confidentiality agreements block us from voicing our concerns, except to the very companies that may be failing to address these issues.”
• There is a risk of dangerous and unexpected outcomes, such as the numerous of instances of social chatbots encouraging their human users to kill themselves along with very specific instructions to do so (Guo, Feb 6 2025).  The company refused to censor the chatbot, reinforcing the validity of Whistleblower concerns about irresponsible development.
• Traditional privacy and data security safeguards have no chance. The tech is unprecedented.

DUDE, THEY ARE CREATING THE DISEASE AND SELLING THE CURE

“GREAT GROUP OF STUDENTS YOU’VE GOT THERE…BE A SHAME IF SOMETHING HAPPENED TO THEIR DATA PRIVACY.

DO YOU NEED ANY PROTECTION?”

DON’T THINK PAST THIS ONE

• It is FAR more naive to assume that the AI developers will give a crap about what we think about AI than it is to assume that we can take a different direction than buying ChatGPT for all our students.
• The google AI budget for next year is $75 billion. The might CSU contract represents 2 one-hundredths of 1 percent of that. That’s not gonna buy us leverage.
• Is this who we want to do business with?
• The WHOLE POINT of AI is that you can’t control it – it will produce solutions that the users can’t anticipate or prevent
• Saying “I want AI but I don’t want it to invade privacy” is like saying “I want to split the atoms, but I don’t want it to make an explosion.” I don’t like it either, but the fact that it sucks doesn’t mean it isn’t true.

​

INESCAPABLE FACT #3: IT IS BEING USED TO DISPLACE WORKERS AND INCREASE WORKLOAD

• A recent Forbes study found that 77% of employees in places that use what is called “A.I.” report that it has increased their workload. This is totally unsurprising. The LLM is generating a huge amount of content really fast that humans have to evaluate and process, so the humans have to scramble to keep up.
• According to this report, companies are already replacing workers with AI. By the way, #2 is Google and #7 is IBM. “In fact, a recent study found that half of CEOs believe they may replace jobs with AI, with the figure rising for C-suite executives.”
• There are ALREADY classes that are entirely run by AI without humans.

​

THREE INESCAPABLE FACTS

• The best available research shows that LLM use by students LOWERS critical thinking ability
• All available evidence shows LLMs are being developed in secrecy, the risks are being hidden and under-estimated in public debates, and the developers ARE NOT open to any outside entity encouraging “responsible” development
• AI is being deployed to CUT the workforce, and the impact of that is to INCREASE the workload of remaining workers.

IMPLICATION #1: THIS COMPLETELY GUTS THE REASONS GIVEN FOR ADOPTING IT

• If it lowers critical thinking skills it makes students dumber. It is ABSURD to suggest that the best way to help our marginalized students is with a technology that makes them dumber. This is NOT equity. This is the opposite.
• There is NO WAY that Google or Microsoft are gonna care about our input IF THEY AREN’T EVEN LISTENING TO THEIR OWN DEVELOPERS.
• In even plainer terms: It is ABSURD to say that giving all of our students ChatGPT is going to give us ANY meaningful input into the direction of AI development.
• It WILL normalize the process and make our students less able to ask the big questions

IMPLICATION #2: WE NEED A TOTALLY DIFFERENT APPROACH

• Invest resources in what REALLY helps students. Research shows the 2 best solutions are (a) smaller classes, and (b) better financial aid. Millea et al., 2018. If we want equity, do the right thing.
• DON’T use it in your classroom, and sign this petition that says you won’t
• Insist that LLMs be used for rote tasks – like travel reconciliations and accounting – BEFORE they are deployed on a large scale in classrooms.
• Query A: “ChatGPT — here are my receipts. Produce a travel reconciliation that is compliant with IRA and system-wide accounting regulations.”
• Query B: “ChatGPT — please identify the core themes of rhetorical theory in Andocides, assess their veracity, and compare them to how they are treated by two different contemporary theorists.”
• If AI really can work, it should be able to handle the lower-level tasks well.
• Focus on digital literacy for our students, educate them about the dangers of LLM use. Do that a long time before encouraging them to use it.

IN CONCLUSION

• The best evidence shows it will make our students dumber
• Repeat that 3 times daily until the fall 2025 semester
• It is totally pollyannish to believe we can influence the AI developers. The most naïve thing we can assume.
• Dear God, let’s try something other than giving it to all the students. We might not have a full formula for how to deal with something everyone is calling inevitable, but at least we can NOT normalize it.
• Please sign the petition or get in touch with me
• It is my deepest hope that when, 300 years from now, Sarah Connor sifts through the smoldering ashes of what was once Cal. State Fullerton, she will find my name on the minutes of a meeting as the last human to resist the AI takeover.

RESOURCES

• Others who think AI will destroy higher education
• A long-form discussion of why AI will take your job, specifically in the CSU
• The full whistleblower link: https://righttowarn.ai/