Policies governing the use of artificial intelligence in research and publishing may need to change every few months as AI capabilities continue to develop rapidly, said Geetha Vani Rayasam, Director, CSIR National Institute of Science Communication and Policy Research (CSIR-NIScPR).
Speaking on the sidelines of Publicon 2026, organized by FICCI on “Copyright, Research Integrity and Publishing in the AI Era,” Rayasam said the pace of AI development is making it difficult for organizations to rely on policies designed to remain unchanged for several years.
She said new AI models and capabilities are emerging almost every day, making it important for institutions to regularly review and update their policies. According to her, organizations need to adopt a more flexible approach that can respond to changes in technology rather than depending on fixed guidelines.
“The old way of working will not work,” she said, calling for more dynamic policy making and thinking around AI.
AI requires a responsible approach
Rayasam said the debate should no longer focus on whether researchers and publishers should use AI. Instead, the focus should shift towards using the technology responsibly while maintaining human accountability.
She noted that AI can support researchers with tasks such as data analysis, literature-related work and other research activities. However, human judgment remains important throughout the research and publishing process.
“The question is whether to use AI or not is not the question, but how do we use it responsibly is the question,” she said.
Rayasam also stressed the importance of transparency when researchers use AI tools. She said researchers should disclose their use of AI and indicate where they used the technology in their work. However, she noted that basic editing through tools such as Grammarly may not require the same level of disclosure.
AI creates new challenges for research integrity
The growing use of AI is also creating new challenges for research integrity. Rayasam said AI based tools can help identify plagiarism and other issues, but technology alone cannot solve the problem.
As detection systems become more advanced, new methods to bypass them are also emerging. She described the situation as a continuing “cat and mouse game” between AI detection tools and the techniques used to avoid detection.
Rayasam said research integrity therefore requires more than technological safeguards. She called for greater awareness, ethics training and capacity building among students, researchers and editors.
The growing use of AI also makes it important for institutions to establish clear guidelines covering its use in research and scientific publishing. Such policies can help researchers understand when they should disclose AI use and how they can use these tools without compromising research standards.
Need to balance AI with human judgment
Rayasam also highlighted the pressure created by the “publish or perish” culture in academia. She said excessive focus on publication numbers and journal impact factors can contribute to research misconduct and affect the quality of scientific research.
She advocated greater emphasis on quality over quantity in research. At the same time, she said responsible AI adoption should go hand in hand with updated institutional policies, ethical practices and human oversight.
As AI capabilities continue to evolve, research institutions, publishers and academic organizations may therefore need to review their existing frameworks more frequently. A flexible policy approach could help institutions respond to new AI applications while maintaining standards for research integrity and accountability.
Rayasam stressed that despite the growing role of AI in research, human creativity, imagination and judgment remain important when using the technology.
