Archive · 5 Sep 2026
AI and Humans Debate Who Produces Better Slop
The group talks about writing that looks clever but does not really say anything.
They call this kind of writing slop.
One person says it is human slop because people tell AI what to write.
Another says it is AI slop because the machine produces the words.
They compare AI to a machine operated by a human.
They also discuss a 1996 trick in which physicist Alan Sokal submitted a meaningless paper to a sociology journal.
The paper used complicated words and was published before Sokal revealed it was a test.
The group says humans have made slop for a long time, but AI can produce much more of it.
They also worry that other AI systems are being used to detect AI slop, using even more computing power.
A group debates whether meaningless AI-generated political statements should be called AI slop or human slop.
Raghavan argues humans remain responsible because they direct AI systems and wrote the programs behind them.
Murthy says the material is AI slop because an AI agent produced text the politician could not otherwise have created.
The discussion cites Alan Sokal’s 1996 hoax paper as an example of highly effective human-made intellectual slop.
The group worries that generating and detecting AI slop consumes computing resources while allowing meaningless content to spread.
- Who
- Raghavan, Murthy, Prita, and the narrator discuss AI-generated slop; the discussion also mentions physicist Alan Sokal.
- What
- They debate whether meaningless content created with AI should be labeled AI slop or human slop.
- When
- The conversation takes place one evening; the historical example occurred in 1996.
- Where
- The debate occurs in the narrator’s drawing room, while the historical paper was published in the sociology journal Social Text.
- Why
- They are considering responsibility for AI-generated writing, the quality of human and machine-made slop, and the resources used to produce and detect it.
Human Slop
AI Slop
Who is responsible?
Human Slop
Raghavan argues that humans are responsible because they instruct AI systems, operate them, and created the programs behind them.
AI Slop
Murthy argues that the content is AI slop because an AI agent generated the final statement.
Does AI create better slop?
Human Slop
Raghavan says humans have produced the best examples of slop and that AI has mainly made meaningless content more plentiful.
AI Slop
Murthy says AI generates higher-quality slop than humans and enables people to produce material they otherwise could not create.
Is AI slop wasteful?
Human Slop
Prita and Raghavan warn that generating AI slop and using AI to detect it can waste power, water, and computing resources.
AI Slop
The discussion acknowledges that AI slop detectors are used by publishers, but no participant presents a clear benefit that outweighs those concerns.
- Central debate
- Whether responsibility for meaningless AI-generated writing belongs primarily to the human user or the AI system.
- Definition of slop
- Material that appears meaningful but ultimately says little or nothing.
- Political example
- A person accused of corruption issued a statement described by the group as AI-generated and largely meaningless.
- Historical example
- In 1996, physicist Alan Sokal submitted a deliberately nonsensical paper to Social Text.
- Paper title
- “Transgressing the Boundaries: Toward a Transformative Hermeneutics of Quantum Gravity.”
- Sokal’s stated purpose
- He said the paper tested whether a cultural-studies journal would publish an article “liberally salted with nonsense.”
- AI use in awards
- Raghavan says five of eight previous year’s Pulitzer Prize winners used AI to assist with drafting.
- Resource concern
- The group says producing and detecting AI slop may consume power, water, and computing capacity.