Why AI Gets Scientific Citations Wrong: A Verification System

Habr AI has published a walkthrough of a thesis project for verifying scientific citations. The prototype accepts PDF and DOCX files, extracts the reference list, checks DOIs, URLs and metadata against Crossref, OpenAlex and other databases, then assigns each entry a reliability status. The core idea is not just to catch broken links, but to separate real sources from convincingly formatted fabrications.

AI-processed from Habr AI; edited by Hamidun News
Why AI Gets Scientific Citations Wrong: A Verification System
Source: Habr AI. Collage: Hamidun News.
Listen to article

Habr AI has presented a system for automatically checking scientific sources. It detects bibliography errors produced by AI models: distorted DOIs, nonexistent authors and broken URLs.

Why do we need a neural network to check scientific citations?

As generative models spread, scientific texts increasingly contain distorted DOIs, mixed-up authors, broken URLs and references to works that do not exist at all.

What does an AI system for scientific citations check?

The system does not merely find the reference list in a document. It checks whether each reference actually exists and whether it can be trusted.

Can a neural network invent a source in a scientific paper?

Yes. As generative models spread, scientific texts increasingly contain distorted DOIs, mixed-up authors and references to works that do not exist at all.

How can you check whether AI made up a reference in a text?

The system described on Habr AI automatically checks every reference: whether it actually exists and whether it can be trusted.

What mistakes does AI make in scientific citations?

As generative models spread, scientific and quasi-scientific texts increasingly contain distorted DOIs, mixed-up authors, broken URLs and references to works that do not exist at all.

Key Questions Answered

Why do we need a neural network to check scientific citations?

As generative models spread, scientific texts increasingly contain distorted DOIs, mixed-up authors, broken URLs and references to works that do not exist at all.

What does an AI system for scientific citations check?

The system does not merely find the reference list in a document. It checks whether each reference actually exists and whether it can be trusted.

Can a neural network invent a source in a scientific paper?

Yes. As generative models spread, scientific texts increasingly contain distorted DOIs, mixed-up authors and references to works that do not exist at all.

How can you check whether AI made up a reference in a text?

The system described on Habr AI automatically checks every reference: whether it actually exists and whether it can be trusted.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…