Klotz paper argues general intelligence may cause extinction

David Klotz's new paper, 'Smart Enough to Go Extinct?' (arXiv:2608.10730, submitted 11 Aug 2026), takes direct aim at the foundational assumption driving AGI development: that flexible, domain-general intelligence is extraordinarily valuable. Klotz builds an evolutionary counter-argument, noting that cyanobacteria have persisted billions of years and horseshoe crabs hundreds of millions without it, while Homo sapiens has survived just 300,000 years and already generates self-created existential risks like nuclear war and climate change. Mass extinction events, he points out, do not preferentially spare cognitively sophisticated species. This leads to what he calls an 'existential risk paradox': general intelligence may be the only biological strategy that produces existential threats to its own possessor.

The paper diverges sharply from mainstream AI safety frameworks. While most accounts trace AGI danger to misalignment—an agent's goals diverging from human values—Klotz locates the risk in structural features of general intelligence itself. Even a perfectly aligned AGI, he argues, would inherit this liability by virtue of being generally intelligent. He then pivots to ethics, invoking deontological duties and the precautionary principle. The uncertainty about intelligence's long-term evolutionary value imposes a duty of caution before engineering AGI or artificial consciousness. If we create a new kind of intelligent being, we are responsible for ensuring conditions in which it can flourish—conditions that, given the track record, we cannot yet guarantee.