Towards More Ethical Research in AI Era: Free Science Approach

The rapid integration of generative Artificial Intelligence (AI) into academic workflows offers unparalleled opportunities for scientific creativity but simultaneously introduces profound ethical risks to research integrity. While the Open Science movement has successfully promoted transparency and reproducibility, its current frameworks were not designed to counter AI-specific challenges such as automated plagiarism, prompt hacking, and the proliferation of low-quality AI-generated publications. Here, we propose the Free Science approach as a comprehensive ethical paradigm to address these emerging threats. This approach extends beyond the principles of Open Science by advocating universal openness, emphasizing the ethical dismantling of systemic biases, and implementing AI-inclusive safeguards. By analyzing the structural limitations of current publishing systems, we demonstrate how the Free Science framework anchored in the pillars of Open Access, Open Data, and Open Peer Review can be adapted to ensure the reliability of scientific outputs in the AI era. Our findings suggest that traditional safeguards are no longer sufficient; instead, a more rigorous ethics-first methodology is required to prevent the normalization of biased or fraudulent knowledge. We conclude that adopting the Free Science approach is essential for preserving the trustworthiness of the global research ecosystem and ensuring that technological advancements enhance rather than undermine scientific progress. This shift not only addresses immediate AI-related concerns but also fosters a more equitable and transparent academic future.