On the ground

The PIPAc consortium, which includes Alysophil, announced in August 2026 a pilot combining synthesis, analysis and AI process control. The press release places systems compatible with pharmaceutical requirements as a future step. The efficiency figures presented are the supplier's.

The exact role of operators. How control is taken back. What remains to be qualified before production. The baselines for the announced gains.

A March 2026 preprint compares two smartphone measurement methods with digitised planimetry, using data from 2023. The automatic method proves less consistent on small wounds. The authors argue for a process of correction and validation by the clinician. The Strasbourg company Pixacare presents its clinical work on its own page.

How clinicians correct the measurement. What the tool has changed in the round or the ward. The version of the device used today.

IHU Strasbourg reports a randomised study conducted with 32 endosonographers on examination videos, followed by four first clinical examinations carried out on 25 August 2026 with mAI Companion. The device analyses a copy of the video stream locally. The study concerns an evaluation on videos; the benefit in routine practice remains to be established.

What the practitioner does with the alert. How the decision is shared. The steps between a first use and routine practice.

The IMARA project, launched in March 2026 by IHU Strasbourg, aims at a platform combining autonomous transport of equipment and preparation gestures learnt by imitation. The operating theatre nurse training institute of the Hôpitaux civils de Colmar is among the partners. The project was selected in the Interreg programme's Offensive Science 2024.

What future theatre nurses learn from this project. The gestures they wish to keep. The timetable for the first trials.

The Hôpitaux civils de Colmar describe an AI-assisted detection module, connected to existing endoscopy equipment, which spots polyps in real time and can predict their nature. The project, launched in 2024 and coordinated by the Haut-Rhin committee of the Ligue contre le cancer, raised €168,000 to equip four Haut-Rhin institutions, including €43,200 for Colmar. The hospital states that the tool does not replace the doctor and announces a comparative study with previous years.

The number of colonoscopies performed with the module, and the status of the comparative study. How the team decides on a biopsy when the screen flags a lesion, and what it does with false alerts. What the four equipped departments share with each other, and what endoscopy nurses take away from it.

TeachIndex lets students give feedback straight after each session. According to the University of Freiburg, the platform analyses this feedback with AI, aggregates it over several sessions and helps lecturers adapt their teaching. Developed since September 2025 by two former medical students, it was the subject of a pilot study at the ENT clinic of Freiburg University Hospital. It is due to be evaluated in the 2026/27 winter semester at 23 study sites. No results from the pilot have been published.

What a lecturer changed in their course after reading the feedback, and how long it took. What exactly the AI does with free-text comments, and who reviews what it draws from them. Whether students are still responding after several weeks, and what anonymity changes in what they dare to write.

The Ortenau district has around 370 volunteer interpreters in six groups, in Achern, Kehl, Lahr, Oberkirch, Offenburg and the Kinzig valley. From January to September 2026 they carried out around 1,788 assignments in more than 50 languages, in public offices, doctors' surgeries, hospitals and schools. According to the district, their role goes beyond translation: explaining the cultural context, spotting a misunderstanding, slowing the conversation down when needed. Its integration officer, Emilie Schleich, considers digital translation a good help in many everyday situations; the statement adds that in important or sensitive conversations, passing on the words is not enough. It does not say whether the volunteers themselves use translation tools.

The moment a volunteer takes out their phone during a conversation, and the moment they deliberately leave it in their pocket. A misunderstanding caused by machine translation that had to be put right. What the training of new volunteers says about translation tools and the confidentiality of what is typed into them.

Asked by a motion to fund intercultural interpreters in hospitals, primary care and outpatient psychotherapy, the Basel-Stadt government replied in a decision of 6 January 2026. It proposes a package of measures: awareness-raising, administrative simplification, a funding pilot and the testing of digital tools, with a report to the cantonal parliament in 2027. It plans to assess translation and interpreting apps in 2026, notably for data protection and reliability. In a survey of psychotherapists, out of 23 responses, 43% use an interpreter and 26% use AI-assisted translation tools; the document notes that the sample is very small.

What makes a clinician choose an app over an interpreter, and the cases where they never do. What is typed into the tool, and how what the patient says is protected. The status of the test announced for 2026: which tools, which practices, what first findings.

The Fédération de Charité Caritas Alsace reported in August 2026 the training of 130 employees of the DITEP Les Tilleuls and La Forge, run with ESEIS. The exercises were linked to confidentiality and to the federation's charter of use. The uses that followed remain to be documented.

What professionals have actually used since. What they chose not to entrust to it. The role of the charter in daily work.

The COLaF project, led by Inria, develops open resources for French and the languages of France. With the Faculty of Languages and the LiLPa laboratory in Strasbourg, it invites contributions to the Common Voice collection of spontaneous speech, including Alsatian. The volumes collected remain to be verified.

How many Alsatian voices have been recorded. Who contributes. What the researchers intend to do with them.

KMØ, in Mulhouse, presents through its Aiguillage programme awareness formats and an Open Lab for testing AI use cases, mentioning local models and the measurement of robustness and productivity. Feedback from a company's actual use remains to be established.

A supported company, its use case, what was kept and what was abandoned.