One place to track AI in biology and medicine.
We follow what AI is doing for living systems: the similarities between the human brain and modern neural networks, the pressing problems medicine still faces, and what AI actually brings to the table, from protein folding to surgery, disability and mental health.
AI for biology and medicine, for a better and humane world.
The film above: Meet AI Biologist, 64 seconds. Music: “Fretless” by Kevin MacLeod (incompetech.com), licensed under CC BY 4.0. Download it in 4K, with the logo and banners, from the brand kit.
Already making a difference.
A few of the results that changed the conversation. Several are early, in the laboratory or on tissue, and we say so.
~200 million
protein structures predicted by AlphaFold and shared openly.
Nobel Prize in Chemistry, 2024
AI-steered
microrobot swarms navigated for targeted drug delivery.
laboratory and model studies, 2025
Autonomous
a robot removed a gallbladder on its own, trained on surgical video.
ex vivo tissue, Science Robotics, 2025
97.5%
of attempted speech decoded by a brain implant for a man with ALS.
NEJM, 2024
51% fewer
depression symptoms in the first randomised trial of a generative-AI therapist.
NEJM AI, 2025
29% more
breast cancers found when AI supports the reading of mammograms.
MASAI trial, Sweden, 2025
More than 1,400 AI-enabled medical devices have been authorised by the US FDA, about three in four of them in radiology.
Our ongoing work.
Radiologists often disagree on where a lesion ends. Instead of training a model to copy one expert, we model the disagreement: our methods emulate many radiologists and give a range of plausible readings.
Patent application filed, 2026.
Sensors that estimate vital signs precisely without penetrating the skin, so monitoring can be gentle, continuous and within reach of more people.
Ongoing work.
We are just getting started, and we are committed to doing much more in this space. Keep watching. optimalstar.org
AI began with ideas inspired by brains. Now it studies the brain, the cell, and disease.
Deep learning traces part of its intellectual history to neurons, perceptrons, and the search for computational models of intelligence. The tools have grown far beyond that origin. They now help researchers read images, predict protein structures, model gene regulation, discover drug candidates, and combine clinical signals that are too large or subtle for manual reasoning alone.
Biology is not just another data domain. It is adaptive, noisy, multi-scale, and deeply coupled: molecules shape cells, cells shape tissues, tissues shape organisms, and organisms respond to environments. That complexity makes the field difficult, but it also makes it one of the most important places for AI to mature.
What we track.
We organize foundational knowledge, open research problems, and practical solutions across the areas where AI is already changing biological and medical research.
Medical diagnosis
Imaging, pathology, ECG, EEG, risk prediction, triage, and clinical decision support systems that can improve accuracy while remaining accountable to physicians and patients.
Drug design
Protein structure, molecular generation, target discovery, virtual screening, toxicity prediction, and the bridge from computational promise to experimentally validated therapy.
Genomics and genetics
Variant interpretation, gene regulation, single-cell analysis, rare genetic disorders, and models that connect sequence, expression, phenotype, and disease.
Personalized medicine
Patient-specific treatment response, multi-omics profiles, digital biomarkers, longitudinal health records, and AI systems that adapt care to individual biology.
Neuroscience
Brain-inspired computation returning to its source: epilepsy, dementia, neural decoding, connectomics, psychiatric disease, and the search for mechanistic understanding.
Infectious disease
Viral and bacterial evolution, outbreak forecasting, antimicrobial resistance, vaccine design, and rapid analysis of emerging biological threats.
Progress will require more than larger models.
AI in biology must confront causality, uncertainty, sparse labels, distribution shift, privacy, clinical safety, laboratory validation, and the difference between prediction and understanding.
A model that scores well on a benchmark is not automatically a cure, a diagnosis, or a biological explanation. The next wave of progress will come from people who understand both computational methods and the biological systems they are meant to serve.
How this site will grow.
Foundations
Clear introductions to biology, medicine, machine learning, statistics, and scientific reasoning for readers who want to participate in AI-based biological research.
Open problems
Living maps of unresolved questions in epilepsy, dementia, genetic disorders, infectious disease, drug response, aging, and healthcare delivery.
Current solutions
Notes on datasets, models, papers, benchmarks, tools, clinical studies, and laboratory workflows that show where the field is succeeding and where it remains fragile.
Learn the foundations. Track the frontier.
AI Biologist is a place to study the ideas, evidence, and open questions needed to work responsibly at the intersection of AI and life sciences. The goal is not hype — it's better research, better healthcare, and longer, healthier lives.