Researcher opts out of all AI-generated text, images, and code interactions

The author states a personal policy of refusing any interaction with stochastic parrots, which includes AI-generated text, images, and code. This means the author will not read websites, ignore emails, or walk away from conversations when AI-generated content is suspected. The author also rejects AI-generated pictures, describing them as hideous and a sign of poor taste. Finally, the author will not run AI-generated code, insisting on human-written code for accountability.

Adopting this policy requires the author to filter out any content that might have been produced by an AI, which can be time-consuming. The author must rely solely on human-produced text, images, and code, potentially losing the speed benefits of AI-assisted drafting or illustration. The author also asks collaborators to send the original prompt rather than the AI output, shifting the burden of expression onto the human. This approach may limit the use of AI-generated summaries or translations that the author would otherwise ignore.

In settings where AI-generated text, images, or code are pervasive, the author’s refusal can lead to isolation from colleagues who rely on those tools. The policy may prevent the author from benefiting from AI-assisted debugging or code generation that could accelerate development. Similarly, AI-generated illustrations that could quickly convey concepts will be dismissed as low quality, potentially slowing visual communication. Consequently, the author’s stance may conflict with workflows that integrate AI as a routine component.

Engineers working with such individuals should provide clear, human-generated alternatives for any AI-produced material they intend to share. Respecting the author’s preference means avoiding the use of AI-generated text as a source or justification in discussions. Offering the original prompt instead of the AI output satisfies the author’s desire to see the human thought behind the request. Maintaining separate channels for human-only communication can help sustain collaboration without violating the author’s boundaries.