Are You a ‘Meat Proxy’? The Rise of Copy-Pasting AI Responses Without Thinking - CNET

Tech > Services & Software > AI

"AI slop" now has a friend. "Meat proxy" is a new way to describe too much reliance on AI content.

Editor / Home Security and Smart Home

It’s increasingly common: You ask a co-worker, associate or buddy a complex question, and they copy and paste an AI-generated response from a chatbot like ChatGPT or Claude, without adding any human input. The internet world now has a name for that sort of person: meat proxy.

The term “meat proxy” seems to have originated in a brief Aug. 3 blog post by developer Niklas Guhn, who complained that he was receiving responses on Slack and other work channels as verbatim output from AI chatbots.

“This is not adding value,” he wrote. “I can talk to Claude myself. It’s going to be faster and I get to control the context. I don’t need a meat proxy in between.”

Social media immediately seized the phrase, with one viral X post declaring, “Babe, wake up, new term just dropped.”

The masses broadened the definition to mean, “A person who forwards AI-generated text, code or other output without reading, understanding or validating it. The person acts only as a relay between the AI system and the intended recipient.” The human is merely the “meat,” acting as an intermediary (proxy) for the machine. Makes perfect sense.

Referring to humans as “meat computers” has also been a growing trend among tech executives — except in this case, it’s about emphasizing the perceived superiority of AI over biological brains. A New York Times article from May noted how that grim vernacular pits humans against machines, portraying people as inefficient, energy-consuming hardware compared to advanced digital superintelligence.

I’ve reached out to linguists and analysts for their opinions on this trend, but in the meantime, it’s easy to see why new slang like “AI slop” and “meat proxy” have resonated. Besides the fact that large language models often get things wrong and produce crappy output, chatbot responses can feel generic, repetitive and lacking effort, especially when they’re flooding spaces that once relied on expertise or authenticity.

Guhn’s issue with meat proxies points to broader concerns about how AI-generated stand-ins dilute human opinions and skills.

In the workplace, for instance, if the person you’re talking to had wanted an automated response, they could have directly asked a chatbot in the first place. Instead, they’re asking you because they want human eyes (and opinions) on the code or project. In the classroom, kids aren’t learning if they’re copying and pasting instead of grappling with the material.

Collaborative, transparent brainstorming isn’t just important for team cohesion and skill building. It also helps prevent AI-generated errors that can creep into content, reducing mistakes and embarrassing fumbles.

Meanwhile, your reading comprehension may take a hit when you’re a mere meat proxy. If you don’t engage with messages and examine them carefully to understand them, you risk losing analytical skills. You may think you’re saving time by taking shortcuts, but frictionless relaying of AI information can undermine how you work with others, and how well you understand news, data reports and even family updates.

It can also lower trust in all kinds of social interactions. Consider apps like Ditto and Winggg, which promise to create AI responses for you on dating apps. That bodes a potential future where, as the head of Bumble has said, it’s all AI models talking to each other.

Politicians, lawyers and other professionals are increasingly getting caught blindly copying and pasting AI-generated material into speeches and briefs. Fabricating citations, reading prompts aloud and other obvious foibles have become easy fodder for ridicule on social media — and serve as a reminder that AI output still needs a human fact-check and edit.

But using AI as a tool for information doesn’t automatically transform you into a meat proxy. Many have been quick to point out the fallacy in assuming everyone leaning on LLMs and chatbots is doing so thoughtlessly.

It’s easy to update your etiquette. Coders employing AI tools at work still need to review that work for flaws or for ways to implement the results more efficiently. Instead of sending an AI-composed message, you could use AI as inspiration for general ideas and then compose your own human-generated response. That both improves your personal voice and makes the engagement more authentic.

In the age of AI, it’ll be essential to develop new habits. But since traps like reposting AI as your homework have become so common, maybe a label like meat proxy is just what we need to raise some awareness.