AI slop
Generic model output that reads long, vague, and obviously templated, usually because the prompt lacked context, examples, and channel-specific constraints.
Michal Juhas · Last reviewed May 2, 2026
Who this is for
Recruiters who hear "AI feels cheap" from hiring managers and want a diagnostic label for bad drafts, not a moral lecture about technology.
In practice
- Set max words per channel: LinkedIn DMs stay short; email gets subject plus scannable body.
- Ban list: phrases like "hope this finds you well" if they signal spam in your market.
- Show one gold example in few-shot prompting before you scale with automation.
Where it breaks
Over-editing until messages sound robotic in a different way, or blaming the model when the brief was empty. Slop is a system problem: prompt, examples, review, and send gates.
Symptom checklist
| Signal | Likely missing input |
|---|---|
| Generic praise | Role and company specifics |
| Wall of bullets | Length cap in instructions |
| Wrong tone | No negative examples |
Related on this site
- Glossary: System instructions, Few-shot prompting, Hallucination
- Blog: How to write better AI prompts, ChatGPT prompts for recruiters
- Membership: Become a member
Frequently asked questions
Why do candidates notice it fast?
They see the same flattery, stacked adjectives, and bullet walls across inboxes. Anything that wastes their time trains them to ignore your brand.
What is the fastest fix?
Add concrete context (company size, stack, location policy), tighten length, and ban phrases you hate. System instructions plus a couple of on-brand examples beat adjectives alone.
Does a better model remove slop?
It can help slightly, but weak prompts produce weak text on any frontier model. Workshops keep returning to few-shot prompting and disciplined Markdown over model shopping.
How does automation change the risk?
Workflow automation scales whatever quality you feed it. Add human review gates until drafts consistently pass your bar.
Is "slop" unprofessional language for executives?
Use the word sparingly externally if it grates; internally it is a useful shorthand for "looks AI-generated." On customer pages, pair the concept with concrete rewrite guidance.
Where can we practice better drafts?
Read How to write better AI prompts, study Markdown for AI packs, and rehearse in a workshop or foundation course.