OpenAI CEO Sam Altman has sparked debate after admitting he increasingly struggles to distinguish between genuine online conversations and content generated by bots or AI models.

Altman described a 'strangest experience' while reading about OpenAI's Codex model, saying comments instinctively felt fake even though he knew the growth trend was real. He said social media rewards, 'LLM-speak,' and astroturfing make communities feel less genuine.

His comments follow an earlier admission that he had never considered the so-called dead internet theory until now, when large language model accounts seemed to be running X. The theory claims bots and artificial content dominate online activity, though evidence of coordinated control is lacking.

Reactions were divided, with some users agreeing that online communities have become increasingly bot-like. Others argued the change reflects shifting dynamics in niche groups rather than fake accounts.