Community-driven and decentralized artificial intelligence in systemic lupus erythematosus: translational insights from the LES_AI initiative

Dear Editor,

Artificial intelligence (AI) is increasingly influencing biomedical research, drug discovery, and clinical care, with machine learning (ML) and deep learning systems capable of integrating multimodal data, identifying therapeutic targets, and supporting real-time decision-making.[1–4] Within rheumatology, these advances are highly relevant due to the complexity, multisystemic involvement, and fluctuating disease activity that characterize systemic lupus erythematosus (SLE).[5] We wish to highlight the translational relevance of LES_AI, a novel AI agent that, to our knowledge, represents the first community-driven and decentralized initiative developed specifically for SLE education and clinical support.

Unlike traditional proprietary AI models, LES_AI was built through a decentralized, community-governed structure in which contributors—clinicians, technologists, and patients—participate in design and content oversight. This development pathway aligns with emerging frameworks for open, transparent, and ethically grounded AI systems. Decentralization also creates opportunities for equitable access, shared governance, and broader representation of patient perspectives, which are particularly important in a disease with substantial disparities in severity and outcomes across demographic groups.[12]

LES_AI integrates natural-language-processing algorithms trained on biomedical literature to synthesize clinically aligned explanations across three domains:

Immunopathogenesis: The agent articulates mechanisms of immune dysregulation, environmental triggers, and epigenetic contributions, echoing current multi-omic and exposomic approaches in SLE pathobiology.[5,6]
Holistic and guideline-consistent management: LES_AI reinforces recommendations consistent with EULAR/ACR guidance, including the foundational role of hydroxychloroquine,[7] appropriate use of immunosuppressants and biologics, and incorporation of evidence-based lifestyle strategies such as anti-inflammatory dietary patterns and structured physical activity, both of which have demonstrated improvements in fatigue, inflammation, and flare reduction.[8–9]
Individualized therapeutic strategies: Its content emphasizes balancing disease control with mitigation of long-term toxicity and promoting adherence, principles central to modern precision-care paradigms.

These strengths illustrate the potential utility of AI agents as adjunctive tools for patient education, clinician support, and knowledge democratization in rheumatology. The webpage’s analysis of AI in medicine underscores similar themes, noting that AI enhances patient outcomes, accelerates drug discovery, predicts toxicity profiles,[4,12,13] reduces administrative burden,[5–7] and improves data-driven clinical decision-making via predictive analytics and CDSS frameworks.[8,10]

However, LES_AI also exemplifies several limitations common across emerging healthcare AI systems. First, formal clinical validation is currently lacking. Without rigorous benchmarking against expert performance, ML-driven systems risk perpetuating bias or producing overconfident outputs that may mislead clinicians or patients.[1,10,11] Second, transparency and explainability remain essential, particularly given the heterogeneity of SLE presentations and the risks associated with opaque “black-box” models.[10,11] Third, interoperability barriers prevent integration with EHRs or structured clinical data streams, constraining real-time utility. Fourth, algorithmic fairness must be prioritized, especially in SLE, where racial and ethnic disparities in organ involvement and outcomes are well-documented.[12]

The decentralized nature of LES_AI may help address some of these concerns by enabling distributed auditing, open governance, and community-based oversight. Nevertheless, safe translation requires structured evaluation pipelines that include accuracy assessment, bibliographic traceability, continuous monitoring for bias, and alignment with emerging regulatory principles for adaptive AI systems.

In summary, LES_AI represents an early but meaningful example of a community-driven, decentralized AI assistant applied to a complex autoimmune disease. Its development highlights both the transformative potential of AI for enhancing patient understanding, supporting clinical communication, and democratizing medical knowledge, and the need for rigorous scientific validation and ethical governance as such systems evolve.

Conflict of Interest

Dr. Erick Adrian Zamora Tehozol is the creator and lead developer of the LES_AI agent and holds digital tokens associated with the project. As such, he may benefit financially from potential increases in the token’s value. These tokens currently constitute the only source of funding for hosting, maintenance, and continued development of LES_AI, as no external or institutional capital is available. The author affirms that the content of this letter was prepared objectively and without influence from financial considerations.

Funding

No external funding was received for this correspondence.

References

Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25:44–56.
Mak KK, Pichika MR. Artificial intelligence in drug development: present status and future prospects. Drug Discov Today. 2019;24:773–780.
Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–589.
Saltz A, Gulko J. AI’s symbiotic impact on patient experience and drug development. LinkedIn Pulse. 2022.
Tsokos GC. Autoimmunity and organ damage in systemic lupus erythematosus. Nat Immunol. 2020;21:663–673.
Rodríguez-Carrio J, et al. The exposome in systemic autoimmune diseases. Autoimmun Rev. 2023;22:103282.
Fanouriakis A, Kostopoulou M, Alunno A, et al. 2019 update of EULAR recommendations for the management of systemic lupus erythematosus. Ann Rheum Dis. 2019;78:736–745.
Gualano B, et al. Physical activity for systemic lupus erythematosus. Arthritis Care Res. 2020;72:1080–1088.
Barbhaiya M, et al. Diet and systemic lupus erythematosus. Curr Opin Rheumatol. 2021;33:576–582.
Amann J, et al. Explainability for artificial intelligence in healthcare. BMC Med Inform Decis Mak. 2020;20:310.
Gerke S, Minssen T, Cohen IG. Ethical and legal challenges of artificial intelligence in healthcare. Nat Med. 2020;26:1327–1334.
Pons-Estel GJ, et al. Understanding the epidemiology and progression of systemic lupus erythematosus. Semin Arthritis Rheum. 2010;39:257–268.
Liu X, et al. Reporting guidelines for clinical trials involving artificial intelligence interventions. Nat Med. 2020;26:1364–1374.

Preprint

Preprint

Community-driven and decentralized artificial intelligence in systemic lupus erythematosus: translational insights from the LES_AI initiative