Identifying AI-generated text through common word patterns: SKIP, 35/100 | Ideas Lab

A reference database of AI writing patterns to help detect machine-generated text.

The pain

Educators, publishers, and content platforms struggle to identify AI-written submissions mixed with human work. False positives damage trust; false negatives undermine integrity.

The gap

Current detection tools (Turnitin, GPTZero) use statistical analysis. A pattern-matching reference could supplement them, but the core problem is that AI models constantly evolve their outputs, making pattern libraries obsolete within months.

Build angle

Start by documenting actual AI tells from current models (Claude, GPT-4, Gemini) with real examples. Sell to academic institutions as a teaching tool first, help students understand what they're writing. Pivot to API integration with LMS platforms if traction emerges.

Strengths

Real, immediate pain point in education and publishing
Low technical barrier to initial MVP (spreadsheet + documentation)
Could generate early revenue from schools willing to pay for detection tools
Educational angle has defensible positioning versus pure detection

Risks

AI models update constantly; patterns become stale within weeks. Maintenance burden is continuous and expensive.
Existing tools (Turnitin, GPTZero, Originality.ai) already do this better with ML. You're competing on pattern curation, not technology.
False positive rate will be high. Flagging legitimate writing as AI damages credibility fast.
Market is shrinking as AI detection becomes table stakes in LMS platforms. Schools expect it bundled, not standalone.
No defensible moat. Anyone can compile patterns; no switching costs once customers realize patterns don't work reliably.
Regulatory risk: detection tools face legal pressure from AI companies and privacy advocates.

Fly Labs Method

Is the pain real, is there a gap, is it the right time, can one person build it.

35

Value Equation

Dream outcome and how likely it feels, against the time and effort it costs.

28

One-Person Business

Curiosity pull, identity fit, and a path from free value to paid for a solo creator.

42

Viral Frameworks

Hook strength, shareability, and how cheaply it can be tested.

38

Builder Lens

Evidence the problem exists, timing, defensibility, and a model that fits on a napkin.

32

Five lenses, one composite. How scoring works

The angle

Start by documenting actual AI tells from current models (Claude, GPT-4, Gemini) with real examples. Sell to academic institutions as a teaching tool first, help students understand what they're writing. Pivot to API integration with LMS platforms if traction emerges.

No market research recorded for this idea yet.

SaaS founders lack a platform for finding partners and organizing joint advertising campaigns with bloggers to reduce costs by 5 times while maintaining the same reach

Other

BUILD

85FL score

Founders waste hours daily on repetitive tasks that AI could automate instantly

Founders spend excessive time on routine tasks that AI could complete in minutes, but no simple, purpose-built tool exists to handle them automatically without complex setup.

Other

BUILD

82FL score

Owners of modern Macs can't launch old games due to complex instructions involving emulators and file conversion

Other

BUILD

79FL score

All Other ideas