Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
134 tokens/sec
GPT-4o
10 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Open and Sustainable AI: challenges, opportunities and the road ahead in the life sciences (2505.16619v1)

Published 22 May 2025 in cs.AI and q-bio.OT

Abstract: AI has recently seen transformative breakthroughs in the life sciences, expanding possibilities for researchers to interpret biological information at an unprecedented capacity, with novel applications and advances being made almost daily. In order to maximise return on the growing investments in AI-based life science research and accelerate this progress, it has become urgent to address the exacerbation of long-standing research challenges arising from the rapid adoption of AI methods. We review the increased erosion of trust in AI research outputs, driven by the issues of poor reusability and reproducibility, and highlight their consequent impact on environmental sustainability. Furthermore, we discuss the fragmented components of the AI ecosystem and lack of guiding pathways to best support Open and Sustainable AI (OSAI) model development. In response, this perspective introduces a practical set of OSAI recommendations directly mapped to over 300 components of the AI ecosystem. Our work connects researchers with relevant AI resources, facilitating the implementation of sustainable, reusable and transparent AI. Built upon life science community consensus and aligned to existing efforts, the outputs of this perspective are designed to aid the future development of policy and structured pathways for guiding AI implementation.

Summary

Open and Sustainable AI in Life Sciences: Addressing Challenges and Forging a Path Forward

The paper "Open and Sustainable AI: challenges, opportunities and the road ahead in the life sciences," presented by a collaborative group of researchers, deeply investigates the pressing challenges, emerging opportunities, and the prospective roadmap for integrating AI within the field of life sciences. It provides a critical analysis of the current state of AI methodologies in biological research, highlighting systemic issues and offering concrete recommendations for fostering sustainable AI development.

Key Research Challenges

The paper identifies three primary areas where significant challenges hinder progress: reusability, reproducibility, and environmental sustainability of AI models. The rapid integration of AI in life sciences has led to an exponential increase in publications, with many models demonstrating non-trivial reusability issues. This problem is largely attributed to deficient documentation of AI methods and insufficient disclosure of model details, making it difficult for subsequent researchers to adapt or redeploy existing models effectively.

Reproducibility is another highlighted challenge, with many AI models failing to provide consistent and reliable benchmarking results due to fragmented methodologies and computational environment discrepancies. This challenge underscores the need for life sciences researchers to embrace practices that ensure reproducible environments and consistent evaluation protocols.

Environmental sustainability emerges as the third challenge, with life sciences AI models consuming substantial computational resources, leading to increased carbon footprints. The paper points out that inefficient AI practices not only waste computational resources but also exacerbate environmental impacts, urging the adoption of Green AI practices.

Recommendations and Implications

To address these challenges, the paper outlines nine specific Open and Sustainable AI (OSAI) recommendations. These include standardizing AI metadata, leveraging AI registries, providing comprehensive training for model deposition, ensuring transparent model documentation, facilitating portable and reproducible model environments, and promoting standardized evaluation datasets. These practices aim to enhance both the scientific and practical aspects of AI research in life sciences, paving the way for more efficient, trustworthy, and sustainable outcomes.

Strategizing AI development with a focus on environmental sustainability is further emphasized through recommendations like implementing Green AI techniques, optimizing hardware choices, and consistently measuring and reporting AI models' environmental impact.

The practical implications of these recommendations are profound, potentially influencing policy development, research funding, and publication standards in life sciences, and by extension, other domains that rely heavily on AI models. Academically, these measures would foster a more collaborative and transparent research landscape, driving technological advancements while considering broader environmental impacts.

Future Directions

The paper calls for continued collaboration among stakeholders, including researchers, policy makers, and funding agencies, to guide the implementation of these recommendations. Moreover, it suggests developing guided pathways, modeled after successful frameworks like the Turing Way, which could serve as toolkits for researchers implementing best practices in AI model development.

In conclusion, the paper provides a comprehensive framework and roadmap for addressing current challenges in AI integration within the life sciences. By emphasizing sustainable and open practices, it lays the groundwork for realizing AI's full potential in driving innovative research while ensuring ecological accountability. Moving forward, the adoption of these recommendations could substantially contribute to the advancement of AI methodologies, ensuring they are both responsibly developed and effectively utilized in addressing complex scientific questions.

X Twitter Logo Streamline Icon: https://streamlinehq.com
Youtube Logo Streamline Icon: https://streamlinehq.com