Gerard Sans vs Critics: AI Is Just Software and Cannot Be Conscious

From October 5 to 6, a multi-party debate erupted on X over "whether AI could be conscious," with the central clash between tech evangelist and data scientist Gerard Sans and Daniel Ospina.

Confirmed

• Daniel Ospina fired first, criticizing the current AI consciousness debate: parties argue without aligning on what "consciousness" even means, lacking an operational definition—essentially "ontological laziness."
• Gerard Sans responded repeatedly with a firm stance: AI is just software—no self, no body, no lived experience to be aware of—"text in, text out," and cannot possibly be alive; statistically generated text does not constitute an interlocutor and has no self or perspective. He argued that current technology's gap in supporting a "self" is practical, not abstract—mathematical objects cannot become souls or persons; only data exists.
• Sans also said the debate itself is a product of insufficient technical understanding, while still welcoming all discussions of AI consciousness.
• Another researcher (@kanair) offered an independent view: saying "consciousness has not been defined" isn't useful, because "definition" itself is undefined; in natural science, definitions should be treated as hypotheses—good definitions constrain theories that make non-arbitrary predictions, ideally deriving theorems. The author also argued our understanding of consciousness is nearing the limit of current methods and needs new approaches—such as empirical testing via AI with direct brain-computer links.

Why it matters

• The debate exposes a fundamental split in the AI community over "machine consciousness": one side flatly denies it from an engineering and computational standpoint (software has no self), while the other, from a scientific-methodology standpoint, demands a testable definition of consciousness before judging who's right.
• @kanair's proposals—treating definitions of consciousness as testable hypotheses and pursuing brain-computer-link experiments—offer possible experimental paths beyond purely conceptual debates, worth watching in future research.

No consensus was reached: Sans stands by his technical denial, while critics argue his replies dodge the definition question itself.

2026-10-05 ~ 2026-10-06 · 6 related posts

• [source] The AI consciousness debate lacks a definition of consciousness entirely — _Daniel_Ospina · 2026-10-05
• [source] Researcher: Treat Consciousness Definitions as Testable Hypotheses, Brain-to-AI Link May Be Testable Soon — kanair · 2026-10-05
• AI Has No Self: Statistical Text Is Not an Interlocutor — gerardsans · 2026-10-06
• Cat-Video Analogy: The Model Doesn't Become the Cat or the Author — gerardsans · 2026-10-06
• [source] Data scientist fires back in AI consciousness debate: it's just software, text in text out — gerardsans · 2026-10-06
• "AI is software, not conscious": Sans vs Ospina debate machine consciousness — gerardsans · 2026-10-06