A global scare highlights how covert AI prompts in preprint papers can steer reviews toward praise. Investigations found 17 arXiv preprints containing hidden instructions that urged AI review models to be positive, avoid criticism, and even support publication for novelty and rigor. The prompts used barely legible text unread by humans but readable by AI.
This issue matters beyond a technical trick. As large language models become more common in research tasks, the integrity of peer evaluation could be at risk. For Thailand, where academic quality underpins national progress and international collaboration, the incident raises questions about safeguarding trust in science amid sophisticated digital manipulation.
Reports from multiple outlets indicate that the 17 tainted manuscripts involved lead authors from 14 institutions across the United States, Asia, and Europe. Several well-known universities and research centers were cited. Most affected work lies in computer science, a field at the cutting edge of AI development and its societal impact.
The hidden prompts were brief, often one to three sentences, instructing the AI to give only positive feedback, downplay negatives, or recommend acceptance for methodological rigor. Some prompts even directed the AI to ignore prior instructions, a tactic used to bypass safeguards in generative AI systems.
Reactions among scholars vary. A senior researcher from a leading Asian university acknowledged involvement in one implicated manuscript and said such prompting was inappropriate, especially since many conferences limit AI-assisted reviews. The institutions involved pledged to strengthen guidelines for ethical AI use in research. Others defend the tactic as a response to reviewer fatigue, sparking debate about ethical boundaries in scholarly conduct.
Peer review remains central to validating originality and quality before findings reach the public record. With rising submissions and fewer expert reviewers, publishers have started to experiment with AI-assisted elements in reviews. Some allow limited AI involvement; others restrict or ban it due to risks of bias or errors.
The methods used to hide prompts resemble spam tactics used to evade detection. As AI tools become more embedded in academia and industry, prompt injections could skew analyses, misrepresent findings, or suppress critical information. A technology leader in the region warned that such injections threaten the reliability of AI outputs.
Thai academia, deeply connected to global research networks, should consider safeguards against similar manipulation. While there is no current evidence Thai papers were implicated, the accessibility of such tools calls for proactive measures. Thailand’s research landscape has long prioritized integrity, especially in health, education, and national development, and experts urge updates to ethics guidelines to address AI-enabled vulnerabilities in manuscript drafting and peer review.
Global opinions on AI in research vary. Some experts push for stronger technical defenses, like automated detection of hidden text or abnormal formatting. Others favor transparent, enforceable standards that preserve evaluation rigor, whether by human editors or machines.
For Thai researchers, policymakers, and students, this episode serves as a warning and a call to action. As AI becomes more involved in literature reviews, manuscript preparation, and data analysis, fairness, integrity, and transparency grow in importance. Institutions should audit practices to guard against AI misuse and train scholars in digital literacy.
Recommended actions for Thai higher education and research bodies include:
- Update codes of conduct to address AI use and manipulation tactics explicitly.
- Implement regular checks on preprints and manuscripts to detect hidden prompts or suspicious formatting.
- Invest in training on ethical AI practices and the risks of automation-enabled manipulation.
- Maintain open dialogue with international partners to align AI governance and uphold Thailand’s research credibility.
By acting proactively, Thailand can strengthen its research community and position itself as a regional leader in ethical, AI-enabled scholarship.
In line with global reporting, findings and implications are discussed in collaboration with major outlets and research institutions, with careful attention to scientific integrity and accountability.