The nephrologist in the present and near future: between algorithms and autonomy

This perspective examines how artificial intelligence (AI) may reshape nephrology over the next two decades while keeping the nephrologist's role central. Prediction models for acute kidney injury and chronic kidney disease progression, and multimodal tools such as KidneyIntelX, will deliver continuous, patient-level risk estimates from electronic health records, biomarkers, imaging, and wearable devices. Large language models (LLMs) are "copilots" for documentation, triage, education, and patient counseling, with potential to reduce administrative burden but also risks of hallucinations, bias, and uneven accuracy. Nephrologists will need new competencies in model calibration, fairness, communication, and ethics to decide when to follow or override algorithmic advice, especially for older, frail, and multimorbid patients. Overall, AI can support proactive, person-centred kidney care only if clinicians help design, govern, evaluate, and critically supervise these emerging technologies within robust, learning healthcare systems.