The AI Slop Scan Hackathon set builders a deceptively hard challenge: stop asking "was this written by AI?" and start measuring whether work actually shows human thinking. Forty-three teams shipped detectors, scanners, and audit pipelines across code review, documentation, marketplace reviews, and general writing. A panel of senior engineers evaluated every submission on detection accuracy, practical usefulness, technical execution, innovation, and presentation.
✯ Top Submissions
- 1st Place — SlopGuard (Team: Team Batman). A content-oversight scoring engine that measures signs of human thinking rather than guessing at AI authorship, reframing the detection problem around information density. Score: 4.19/5.0 across 8 judges.
- 2nd Place — Signal-OSS (Team: Signal-OSS). A zero-tolerance slop scanner for code review that measures the actual information density of pull requests, commit messages, and code comments instead of running an authorship classifier. Score: 4.15/5.0.
- 3rd Place — SlopLens (Team: SlopLens labs). A three-layer hybrid detection engine that scores any text for information density, filler, and naturalness, shipping as a web app, Chrome extension, REST API, and CLI. Score: 4.14/5.0.
✯ Excellence Tier
- Showreceipts (BharatShowreceipts-DELTA) — introduces DELTA, a metric that separates descriptions which restate a diff from those that explain why an approach was chosen. Score: 4.10/5.0.
- SlopBlock (Blockers) — an on-device Windows app that detects and hides AI-generated text, images, and video in real time as you browse, with no cloud or telemetry. Score: 4.01/5.0.
- BS-Meter (Momo) — a multi-domain detector scoring text across six domains using twelve deterministic signal categories. Score: 3.95/5.0.
✯ Standout Innovations
- Papyrus (OneAbove) — audits the reference layer rather than authorship, verifying that citations exist and scoring evidentiary claim alignment.
- Review Constellation (Nebula) — an explainable review-fraud platform that detects coordinated behavioural patterns instead of guessing at AI authorship.
✯ Evaluation Panel
Submissions were scored on a weighted five-criterion rubric (detection accuracy, practical usefulness, technical execution, innovation, and presentation). The senior panel included:
- Amr Arqoub — Innovation Director of Technology & Partnerships, Freshpet
- Iuliia Kozlova — Lead Software Testing Expert and CNCF Kubestronaut
- Myroslav Mishov — Lead Enterprise Architect and CNCF Kubestronaut
- Sergii Demianchuk — Senior Software Engineering Technical Leader, Cisco Systems
- Sushil Choubey — Principal Supply Chain Manager, Amazon
