Scientific Literacy and Security Challenges in the Age of Artificial Intelligence

While artificial intelligence technologies enhance access to and understanding of scientific information, they create new vulnerabilities regarding research reliability, data manipulation, and fake citations. According to Giulia Stefenelli, scientific literacy, verification processes, and publisher responsibilities must adapt to this change.

Ease of Access to and Understanding of Scientific Information

Artificial intelligence plays a significant role in clarifying the scientific evidence necessary to make informed decisions on public policy, environment, and health issues. Open-access publishing and generative artificial intelligence largely remove the barriers to reading and understanding scientific literature.

Falling Production Costs and Verification Challenges

The ability to produce content quickly and at a large scale creates an imbalance between the time and expertise required to verify data and references. While generating content with artificial intelligence becomes cheaper, verifying this information remains a costly and laborious process.

Security Vulnerabilities Created by Artificial Intelligence

The technology can manipulate research components entering the scientific record and distort original research through selective presentations. Fake citations in language models and suspicious references in biomedical articles are on the rise.

Transparency and Shared Responsibilities

Reliable research requires authentic data, transparent reporting, and accountable individuals. The verification process must not only detect artificial intelligence-generated texts but also examine whether the study was genuinely conducted and whether the results are logical. Publishers, researchers, and artificial intelligence providers share responsibility regarding transparency.