AI;DR: The Emergence of a New Social Contract for AI-Generated Content

The AI;DR Policy: Refusing Unedited AI Slop

AI;DR (AI; Didn't Read) is a social response to the proliferation of raw, unedited AI-generated text, where the recipient refuses to read content if the sender did not bother to review or edit it. This policy mirrors the "TL;DR" (Too Long; Didn't Read) phenomenon but focuses on the quality and intent of the communication rather than its length. The core premise is that sending unfiltered AI output is a sign of intellectual laziness and a failure to respect the recipient's time.

In professional and personal communications—such as Slack discussions, newsletters, and social media—the presence of "AI-isms" (generic, flowery, or overly verbose language) signals that the sender may not fully understand or stand behind the prose. Consequently, the recipient feels no obligation to invest the effort required to parse the content.

The "Meat Proxy" Problem and Externalized Costs

Sending raw AI output transforms the human sender into a "meat proxy," where the person acts merely as a conduit for a machine without adding human judgment or verification. This creates a dynamic where the sender externalizes the cognitive cost of reviewing the material onto the recipient.

Key issues identified with this behavior include:

• Abdication of Responsibility: When a sender provides a wall of AI text, the recipient becomes the first actual human reviewer, forced to weigh the merit and plausibility of the claims that the author should have vetted.
• Loss of Provenance: Using AI as a middleman obscures the origin of the ideas and can lead to miscommunication or delayed clarity.
• Erosion of Trust: Writing is a primary vessel for building trust. When a reader suspects the writer hasn't actually read their own output, the social contract of communication is broken.

Impact on Software Engineering and Technical Documentation

The proliferation of AI-generated content is creating a "post-readability" environment in technical codebases and peer reviews. Engineers are reporting a surge in AI-generated documentation and comments that add volume without adding value.

Specific technical frictions include:

• PR Noise: Pull Requests (PRs) are increasingly filled with AI-generated commit messages and documentation that read like formal bills rather than concise summaries, forcing reviewers to work harder than the authors.
• Low-Nuance Technical Claims: There are reports of AI-written technical articles making bold claims about complex architectures (e.g., PCIe devices over TCP/IP) while completely omitting critical technical nuances like DMA, interrupts, or IOMMU.
• Codebase Bloat: AI-generated comments are becoming a "second documentation system" that often consists of noise, describing what the code is doing (which should be self-explanatory) rather than why it is doing it.

Proposed Alternatives to Raw AI Output

To maintain professional etiquette and communication efficiency, several alternatives to sending raw AI text have been proposed:

• Share the Prompt, Not the Output: Instead of sending the generated text, send the prompt used. This conveys the exact intent and information the sender wants to communicate without the "flowery language" and guesses added by the LLM.
• Own the Session: In collaborative tasks, provide the model and prompt so the recipient can steer the AI session themselves, rather than receiving a static, potentially irrelevant report.
• Human-in-the-Loop Editing: Use AI for sourcing ideas or outlines, but ensure the final prose is edited by a human to remove generic AI mannerisms and ensure factual accuracy.

Counterpoints and Nuances

While the AI;DR sentiment is widespread, some argue that the focus on how content is produced can overshadow its actual utility.

"I personally don't care whether a piece of writing was AI-generated as long as it's useful or insightful... statistically speaking 90% AI-generated pieces are a waste of readers' time. It's a good strategy to simply skip them."

Other considerations include:

• Accessibility for Non-Native Speakers: For those communicating in a second language, AI can be a vital tool for translation and clarity. However, critics note that even in these cases, "broken English" that conveys a genuine human idea is often preferred over polished but empty AI prose.
• The "AI Tell" Paradox: As AI is trained on high-quality human writing, the markers of a good writer (such as the use of em-dashes or specific technical terms) are increasingly being misidentified as "AI tells," leading to innocent writers being accused of using LLMs.