Which hospitals, medical foundations, and private and public health research institutes in the Paris region carry out real, deep machine-learning, AI and data research projects for the medical sector?
An evidence-backed web scan (current to 24 October 2026, roughly 4,700 search results reviewed) found 94 organizations in Paris or Île-de-France with substantive, original medical AI research — medical foundation models, radiology and pathology computer vision, neuroimaging, clinical NLP, multimodal oncology, drug discovery, federated learning, digital twins, omics and causal health-data science. Inclusion required evidence of original research, not just adoption of off-the-shelf AI tools. Coverage caveat: no authoritative registry exists for this cross-cutting universe, so 94 is the complete evidence-backed set found by this scan, not a guaranteed census. Categories are evidence-based classifications; some rows rest on secondary pages or professional profiles — each row keeps its source link so you can judge it.
Each dot is one organization, sized equally; bar length is the category count from the precomputed overview. Hover a category bar for its evidence depth.
Organizations by labelled location area (no coordinates supplied; grouped from the directory's city and site labels)
Organizations whose stated focus mentions each method or clinical theme (one organization can count in several)
Search any medical area, method or name; filter by category. Source mentions indicate evidence depth in this scan, not research quality. Rows with a single mention are kept.
| Organization | Category | Location | Representative project / team | Research focus | Mentions | Source |
|---|
Method: evidence-backed web scan, current to 2026-10-24, of about 4,700 search results across hospital, public/nonprofit and private-sector query families; 94 organizations in Paris or Île-de-France kept after merging spelling variants, one row per organization with its own source URL across 36 source hosts (16+ independent hosts overall). "Mentions" counts source mentions found for the organization in this scan — an evidence-depth cue, not a quality ranking. Category totals (39/30/22/3) come from the precomputed overview. Research-focus text is shortened in the table for space; location areas are grouped from supplied labels, 7 rows carry no location label.