Skip to content
Submit incident
AII-1002-44693
Well documentedThree or more independent publishers report the core facts.

AI-generated fake papers found on Google Scholar

A Harvard Kennedy School study identified 139 AI-generated research papers appearing on Google Scholar, many containing characteristic phrases from ChatGPT. The papers focused on controversial topics like health and climate change, raising concerns about the manipulation of scientific evidence and public trust.

9 September 2024 Β· Google Β· Misinformation and synthetic media
9 September 2024
AII-1002-446934 sources ↓
Quick facts
Category
Misinformation and synthetic media
Country
United States
Source
Newsweek
Date
9 September 2024
Drafted by
Raidu Incident Detection

What happened

A study published in September 2024 by the Harvard Kennedy School's Misinformation Review identified 139 research papers suspected of being generated by AI tools on Google Scholar. The papers contained phrases commonly produced by ChatGPT or similar programs, such as 'I don't have access to real-time data' and 'as of my last knowledge update.'

Most of the AI-generated papers were found in non-indexed journals and working papers, though some appeared in mainstream scientific journals and conference proceedings. The papers often focused on policy-relevant topics including health, environmental issues, and computing technology.

Reported impact

The study warned that AI-generated papers could facilitate 'evidence hacking'β€”the strategic manipulation of society's knowledge baseβ€”and pose a threat to public trust in science and evidence-based decision making. The authors noted that generative AI makes it easier, faster, and more convincing to create and spread questionable research at lower cost than traditional methods.

Index analysis

The researchers analyzed scientific papers from Google Scholar for signs of GPT use, identifying 139 'questionable' papers as regular results on the platform. More than half of these papers focused on health, environmental issues, and computing, topics susceptible to disinformation.

Governance controls

Researchers acknowledged that AI-generated papers could be removed but warned that doing so might fuel conspiracy theories. The authors cited the rise of anti-vaccine movements during COVID-19 and ongoing climate change denial as examples where retracting erroneous publications increased rather than decreased following of those movements.

Classification

Category
Misinformation and synthetic media
Industry
Healthcare, drugs and biotechnology
Country
United States
Harm status
Potential harm
Intent
Unknown
Timing
Unknown
Harm types
Public interest, Reputational
Harmed
General public
Failure modes
Fabricated citation, Hallucination
AI systems
ChatGPT

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Newsweek
AI-generated junk science is flooding Google Scholar, study claims
9 September 2024
PYMNTS.com
AI-Generated Junk Science Research a Growing Problem, Experts Say | PYMNTS.com
10 September 2024
ZDNet
Harvard study finds AI-generated papers on Google Scholar - why it matters
11 September 2024
Business Insider
Fake GPT-written studies are flooding Google Scholar. Here's why taking them down could make things worse.
12 September 2024