"Rather Fail with Honor than Succeed by Fraud"

A Conversation with StealthGPT Creator Jozef Gherman

StealthGPT can write academic papers. It can also write peer-reviewed papers and Ph.D. theses. And it can write them in such a way that they are undetectable by AI.

If the university system was as pure as the driven snow and administrators fired plagiarists, and fired people who fabricated data (like the University of Minnesota’s famous Alzheimer's researcher Sylvain Lesné who had one of the most cited peer-reviewed papers retracted—which cost billions of dollars and countless lost hours)… if the university system was acting ethically and adhering to its own rules, then they would have a leg to stand when criticizing StealthGPT. But they are not acting ethically and playing by their own rules.

Universities could criticize StealthGPT if faculty, administrators, and staff weren't guilty of egregious ethical violations and were then shielded from the consequences of their behavior. But universities are almost universally guilty of egregious ethical violations so by what basis can they criticize a tool that enables students to cheat when their own faculty is cheating and they choose to ignore, deny, attack those who point out the cheating, and don't fire conspicuously guilty faculty?

StealthGPT is a tool that has the potential to be incredibly disruptive. In fact, more than incredibly disruptive, perhaps utterly destructive of the whole academic enterprise, including pedagogies, take-home exams, take-home essays, the possibility of creating students who can think and write, etc. We are faced with a problem of how our institutions will deal with a tool that enables large numbers of students to cheat while the universities turn a blind eye to their faculty cheating.

We also have the additional problem that faculty are using StealthGPT to cheat because now they have the tools to publish peer-reviewed papers that they did not write—the papers were AI-generated. Faculty could use AI-generated papers to leverage themselves into tenure, that is, jobs for life.

Our present system cannot survive this. It is irredeemably corrupt and utterly beholden to invasive, parasitic ideologies. It is completely lacking in moral authority and public trust. We need to create new systems that take into account these radical technologies and are divorced from corrupt relics of the past. It is currently unclear if we are capable of doing this at scale, and the consequences of failure will be incalculable.

Find Jozef Gherman, Founder of StealthGPT on X.