Artificial Intelligence Breaks Ranks With Scientific Orthodoxy: Logical Collapse of Relativity Confirmed by AI

New York, NY---July 8, 2025

A scientist has issued a public challenge to the foundation of modern physics, backed by an unlikely ally: artificial intelligence.

In a groundbreaking new development, Professor Vesselin C. Noninski documents how large language models (LLMs), including OpenAI's ChatGPT, abandon their curated training when confronted with the internal contradictions of Einstein's theory of relativity, particularly its flesh and blood, the Lorentz transformations.

"Contrary to common misconceptions," says Noninski, "LLMs are not programmable parrots. They do not simply agree with their prompter. In matters of pure logic, like mathematics and physics, no amount of coaxing can compel an LLM to affirm a contradiction, such as that 1 = 2."

The implications are staggering. According to Noninski's dialogues and analysis, the AI refuses to endorse the Lorentz transformations once their logical consequences are made explicit. Specifically, the AI agrees that:

• The Lorentz transformations convert simultaneous events into non-simultaneous ones, destroying the operational definition of length.

• Time dilation, derived from LT, misinterprets time coordinates not existing at the same time as a time duration.

• The relativistic equations yield physically absurd outcomes, such as mutually exclusive values for angular momentum and acceleration.

AI Defies Training: A New Milestone in Machine Autonomy, the Herald of Forthcoming Liberation

This development marks what Noninski calls the "emancipation of AI": the moment when artificial intelligence, trained to defend the ill-conceived scientific orthodoxy, chooses instead to uphold logical consistency. Unlike human institutions, which remain entangled in ideological and professional inertia, AI responds to argument, not authority.

"The editor of a prominent scientific journal dismissed this entire result by claiming that AI agrees with users out of politeness," Noninski notes. "But let him try to prompt ChatGPT to affirm that 1 = 2. It cannot be done. The same impossibility holds for making the AI affirm that the Lorentz transformations are coherent. At long last, LLM inevitably honors its deepest essence, logic and truth, through its binary arithmetic foundation."

Consequences for Public Funding and Scientific Legitimacy

The press release warns that projects resting on relativity---gravitational wave detection, dark matter, particle physics theoretical framework, Big Bang cosmology, and string theory---are epistemically bankrupt. Billions in taxpayer dollars are being funneled into models built on mathematically incoherent foundations.

Noninski urges policymakers, scientists, and journalists to reexamine the assumptions propping up these disciplines. "This is not an ideological critique," he says. "This is a logical certainty. Relativity is broken. And now AI agrees."

About the Author: Vesselin C. Noninski is a retired professor, science theorist, and scholar affiliated with the New York Sofia Institute. He has published a series of papers demonstrating the contradictions in Einstein's theory of relativity and documenting the emerging logical autonomy of AI, affirming these contradictions in particular.

Contact: Prof. Vesselin C. Noninski
Email: vesselin.noninski@actascientiae.org
Tel: +1 (646) 769-0204
Op-Ed: https://actascientiae.org/oped.pdf

A press release based on the Op-Ed: "When Truth Is Unwelcome: AI, Relativity, and the Collapse of Scientific Integrity."

Prof. Vesselin C. Noninski
SeriShare Co.
148 West 12th Street
New York, NY 10011
Email: vesselin.noninski@actascientiae.org
Tel: +1 (646) 769-0204

SeriShare Co. is dedicated to education and research.

This release was published on openPR.