Is AI-based fall detection high-risk under the EU AI Act? This article explains how MOBOTIX NurseAssist is classified, why, and what this means in practice for resellers, system integrators and care providers.
Status: October 2026
Why the AI Act matters for NurseAssist
NurseAssist brings AI-based video analytics into patient and resident rooms, so partners and customers rightly ask how the EU AI Act applies. The short answer: used as intended, NurseAssist generally falls outside the Act’s high-risk categories.
The EU AI Act in brief
The AI Act (Regulation (EU) 2024/1689) regulates AI by risk. Some practices are prohibited, listed high-risk systems face strict requirements, systems such as chatbots carry transparency duties, and most other AI faces only general provisions such as AI literacy.
Prohibitions and AI literacy have applied since February 2025, transparency duties since August 2026. The Digital Omnibus on AI (Regulation (EU) 2026/1744) moved the high-risk requirements to December 2027 (Annex III) and August 2028 (Annex I).
How is MOBOTIX NurseAssist classified?
Intended purpose
NurseAssist recognises predefined situations in a patient or resident room and notifies care staff, typically via a nurse call system. Both MOBOTIX c ONE NurseAssist and MOBOTIX c71 NurseAssist detect falls and bed occupancy or bed exit. Licensed extensions add events such as sitting on the edge of the bed or entering and leaving the room, and c ONE NurseAssist also covers getting up from bed or a chair and changes in lying position. The intended purpose is identical, so one assessment covers both.
Why NurseAssist is not high-risk
Article 6 offers two routes to high-risk status. The first concerns products, or their safety components, that are regulated by Annex I legislation such as the Medical Device Regulation and require third-party conformity assessment. NurseAssist is not classified or intended as a medical device, so this route does not apply.
The second covers the use cases in Annex III. For camera-based AI, the relevant ones are remote biometric identification, biometric categorisation by sensitive attributes, and emotion recognition. NurseAssist does none of these: it does not establish who someone is, sort people into categories or infer feelings or intentions. A fall or a person sitting on the edge of the bed is a physical situation. The AI Act explicitly separates emotions from physical states such as pain or fatigue, as do the Commission’s draft high-risk guidelines (final version expected by the end of 2026).
Annex III’s healthcare-related entries, eligibility decisions by public authorities and emergency patient triage, do not fit either: NurseAssist does not decide who receives care or in what order.
Article 50’s transparency duties are not triggered. They target conversational AI such as chatbots, AI-generated content and deepfakes, and emotion recognition or biometric categorisation.
When the assessment changes
Classification follows the intended purpose. Under Article 25, anyone who changes the intended purpose of a non-high-risk AI system so that it becomes high-risk is treated as its provider, with all provider obligations. Examples include using NurseAssist events to evaluate staff performance (an employment use case in Annex III), linking it with facial recognition to identify residents, or promoting it for diagnosis or fall-risk prediction, which could make it a medical device.
What this means for partners and care providers
MOBOTIX defines and documents the intended purpose and limitations, provides product and installation information, and is responsible for the regulatory requirements that apply to its products.
Distributors and resellers have no specific AI Act duties here, as the Act’s distributor obligations apply only to high-risk systems. As good practice, present NurseAssist accurately: as a support tool, not as a diagnostic, identification or staff-monitoring tool. Promoting other uses can shift legal responsibility.
System integrators should install NurseAssist according to MOBOTIX documentation, including the scene requirements. We recommend agreeing customer-specific changes to what the system detects or how its events are used with MOBOTIX in advance.
Care providers are deployers under the AI Act, and one legal obligation applies: Article 4 requires them to take measures to support the AI literacy of staff who use the system. Since July 2026, the text states that no specific level has to be guaranteed. A briefing on what NurseAssist detects, its limits and how to respond to alerts is a practical starting point. We also recommend defined alert-handling processes. NurseAssist supports professional judgement; care decisions remain with qualified staff.
AI Act and GDPR: two separate questions
The two regulations apply independently, and a non-high-risk classification says nothing about GDPR compliance. Video (analytics) in residents’ rooms can process personal and health-related data. As controller, the care provider must establish a legal basis, inform residents and staff, involve employee representatives where required and, in many cases, carry out a data protection impact assessment.
FAQ
Does a reseller need AI Act certification? No. There is no such certification. Conformity assessment applies only to high-risk systems and is the provider’s task.
Does a care facility need a formal AI Act risk assessment? Not for standard use. Risk management and fundamental rights impact assessments are tied to high-risk systems, although a GDPR impact assessment may be needed.
What about existing c71 NurseAssist installations? The assessment depends neither on platform nor on installation date. Beyond AI literacy, installations used as intended need no AI Act-specific action.
Conclusion
NurseAssist recognises physical situations and leaves decisions to people. Used as intended, it generally sits outside the AI Act’s high-risk regime, and what remains is manageable: accurate communication, documented installation, informed staff and sound data protection. Planning a non-standard use? Involve MOBOTIX early.
Sources
- Regulation (EU) 2024/1689 (AI Act), Official Journal
- AI Act, consolidated version as of 27 July 2026
- Regulation (EU) 2026/1744 (Digital Omnibus on AI)
- European Commission AI Act Service Desk
- Draft Commission guidelines on the classification of high-risk AI systems
- Commission guidelines on transparency obligations under Article 50
This article provides general regulatory information, not legal advice. Applicable requirements depend on the specific use case and legal context.