Virtual And Digital Health Digest – August 2026

Welcome to the latest installment of Arnold & Porter’s Virtual and Digital Health digest. This digest covers key virtual and digital health regulatory and public policy developments during July and early August 2026 from the EU and UK.

During July 2026, regulators in the EU and UK continued to focus on the practical implementation of AI and digital health regulation.

In the EU, the Digital Omnibus on AI introduced the first substantive amendments to the AI Act, while policymakers and industry groups turned their attention to how AI requirements will interact with existing frameworks. This included new principles on the use of AI in Joint Clinical Assessment dossiers under the Health Technology Assessment Regulation, the European Federation of Pharmaceutical Industries and Association’s (EFPIA) call for trusted AI testing environments and proportionate AI governance in its response to the Biotech Act consultation, and further debate on the treatment of software and AI-enabled technologies in the ongoing Medical Device Regulation (MDR)/In Vitro Diagnostic Regulation (IVDR) revision. Together, these developments reflect a growing focus on ensuring that AI-specific requirements integrate effectively with existing regulatory frameworks.

In the UK, regulators published a series of guidance documents addressing questions about when AI-enabled products qualify as medical devices and how they should be regulated. New Medicines and Healthcare products Regulatory Agency (MHRA) guidance on ambient voice technologies and updated guidance on digital mental health technologies provide greater clarity on the medical device boundary, while additional outputs from the AI Airlock program highlight the challenges associated with qualification, intended purpose, and lifecycle management of AI systems. These developments suggest that regulators are increasingly focused on addressing practical implementation questions.

Regulatory Updates

EU Adopts Digital Omnibus on AI to Simplify Implementation of the AI Act. The EU has adopted the Digital Omnibus on AI (Regulation (EU) 2026/1744), introducing targeted amendments to the AI Act aimed at reducing regulatory burden and improving alignment with sector-specific legislation. Key changes include clarifying the scope of high-risk AI systems and safety components, expanding AI regulatory sandbox provisions, strengthening the powers of the EU AI Office, and clarifying the interaction between the AI Act and other frameworks, including the Cyber Resilience Act. The Digital Omnibus on AI entered into force on July 27, 2026, with staggered application dates for various amended EU AI Act provisions.

European Parliament Publishes Draft Report in Response to the European Commission’s Proposal to Amend Regulation (EU) 2017/745 (MDR) and Regulation (EU) 2017/746 (IVDR). The draft report shows broad support for the European Commission’s reforms, while suggesting further targeted amendments. The European Parliament proposes additional measures to support access to “niche” medical technologies and to streamline classification and conformity assessment processes. For digital health companies, it includes further revisions to the proposed software classification rules and supports more structured pre-submission engagement with notified bodies, although it is not expected to significantly expand the circumstances in which software can qualify as a Class I device. Read our July 2026 BioSlice Blog for more details on the proposed amendments.

EFPIA Publishes Response to the EU Biotech Act Consultation. Following its position paper last month (see our July 2026 Digest), EFPIA has now published its response to the European Commission’s consultation on the proposed EU Biotech Act. The submission condenses and reinforces its key recommendations, which include measures to support data- and AI-driven innovation. In particular, EFPIA supports clarification of the processing and reuse of clinical trial data, greater harmonization of GDPR requirements, the introduction of regulatory sandboxes, and the development of trusted AI testing environments. The submission also emphasizes that AI and data governance should be proportionate, risk-based, and aligned with existing legislation, while avoiding unnecessary duplication.

Team-NB Proposal for Risk Adaptive Surveillance System. Team-NB has published a position paper setting out a proposal for revisions to the EU IVDR to establish a risk-adaptive surveillance system for In Vitro Diagnostics (IVDs), which would include software- and AI-based IVDs. Under the proposal, manufacturers would initially be subject to enhanced surveillance following certification, with the possibility of reduced oversight over time where they demonstrate sustained compliance, effective quality management systems, and strong post-market performance. The proposal is intended to support the European Commission’s ongoing work to move away from fixed certificate validity periods and towards a system of periodic reviews, with the intensity of notified body oversight determined by the risk profile and compliance history of the device and manufacturer.

European Commission Adds New Exemptions to Portable Battery Removal Requirements. The European Commission has adopted a delegated act adding six new product categories to the list of exemptions from the Batteries Regulation requirement that portable batteries be removable and replaceable by consumers. The new exemptions include certain wearable devices (such as smartwatches and fitness trackers), meaning that batteries in those products may instead be removable and replaceable by independent professionals. The European Commission has updated its guidance on battery removability and replaceability to assist manufacturers in applying the new derogations. The delegated act will now be scrutinized by the European Parliament before it can enter into force.

MDCG Publishes Position Paper on UDI Assignment Between Manufacturers and Distributors. The Medical Device Coordination Group (MDCG) has published a position paper clarifying responsibility for assignment of Unique Device Identifiers (UDI-DIs) under the MDR and IVDR. The document addresses instances where distributors have assigned their own UDI-DIs to products marketed under their brand name, creating uncertainty about how devices should be identified and registered in EUDAMED. The position paper notes that responsibility for UDI assignment and registration remains with the manufacturer. While devices marketed under different brand names may have different UDI-DIs, those identifiers should remain linked to the manufacturer in EUDAMED.

HTA Coordination Group Publishes Principles on AI Use in Joint Clinical Assessment Dossiers. The European Commission has published new General Principles on the use of AI in the preparation of the Joint Clinical Assessment dossier under the EU Health Technology Assessment Regulation. The principles acknowledge that AI may be used in evidence synthesis and dossier preparation, but make clear that its use does not alter the legal, methodological or evidential requirements applicable to Joint Clinical Assessment (JCA) submissions.

UK Government Reports on Progress of Life Sciences Sector Plan. The government has published a progress report on the delivery of the Life Sciences Sector Plan, launched in July 2025 (see our July 2025 BioSlice Blog on the plan’s publication). The update reports progress on four headline targets: (1) increasing investment in commercial research and development; (2) improving access to “scale-up” capital; (3) accelerating patient access to medicines and medicinal tech; and (4) increasing foreign direct investment. The report highlights the government’s efforts to accelerate the adoption of digital health and medtech innovation through the National HealthTech Access Programme, value-based procurement reforms, and streamlined NHS adoption pathways. It also emphasizes investment in digital health infrastructure, including the new Health Data Research Service and regulatory initiatives such as the MHRA’s AI Airlock program. Together, these measures are intended to support faster patient access to innovative technologies and strengthen the UK’s position as a leading location for health technology innovation. 

MHRA Publishes Guidance on Ambient Voice Technology-Enabled Products. The MHRA has published new guidance on ambient voice technology (AVT)-enabled products. AVTs are AI-powered tools that automatically capture and convert spoken words into text and/or other outputs. Increasingly, AVTs are GenAI-based, with powerful, general-purpose Large Language Models underlying. The guidance provides practical examples of when an AVT-enabled product will not qualify as a medical device, including: (1) ambient scribes intended to provide a transcript or summary of a clinical conversation between a clinician and a patient; (2) AVT products that format information from an encounter for a clinician to review, edit, and confirm; and (3) AVT products that suggest possible clinical codes for clinician review based on a clinical conversation. The guidance also provides examples of AVT products that will qualify as medical devices, such as AVT products that can provide “generated insights” that suggest diagnoses or treatment options. There is also guidance on the risk classification of AVT products that qualify as a medical device. 

MHRA Updates Guidance on Digital Mental Health Technologies. The MHRA has updated its guidance on the qualification and classification of digital mental health technologies (DMHT), which is intended to help manufacturers determine whether their products qualify as software as a medical device (SaMD). The revisions include additional information on the circumstances in which a technology would not be a medical device. For example, the guidance explains that some digital technologies will be made up of various software modules where some qualify as SaMD and others do not. Manufacturers can choose to regulate the DMHT modules that qualify as SaMD as separate products or all the modules as one product. The guidance has also been updated for consistency with the new guidance on AVTs discussed above. 

MHRA Updates Guidance on Registration of Medical Devices. The MHRA has updated its guidance on registering medical devices placed on the market or put into service in Great Britain. The guidance clarifies the MHRA’s expectations with regard to registration of devices that are “put into service” but not “placed on the market,” in that MHRA registration is encouraged for such devices, though this is not legally required. New guidance also provides practical examples of when a device should be considered to be “put into service.” 

UK Parliament Launches Work on Regulation of AI in Drug Discovery. The Parliamentary Office of Science and Technology (POST) has announced a new project examining how AI models used in drug discovery are evaluated and regulated in the UK. The work will explore the growing use of AI to design novel molecules, predict drug behavior before clinical testing, and support decisions on which drug candidates progress into development, alongside questions about validation, bias, reliability, and regulatory oversight. POST notes that the global AI drug discovery market is estimated at approximately $1.7 billion in 2024 and could exceed $8 billion by 2030. Stakeholders have been invited to contribute evidence until September 18, 2026, with publication of the final outcome expected in December 2026.

MHRA Publishes Case Studies From AI Airlock Sandbox Phase 2. The MHRA has published case reports from Phase 2 of its AI Airlock Sandbox, providing practical insights into regulatory challenges faced by developers of AI as a medical device. The case studies informed the recommendations set out in the AI Airlock Sandbox Phase 2 program report (as reported in our July 2026 Digest) and cover a range of topics, including intended purpose and validation, AI-powered in vitro diagnostic devices, regulatory intelligence, post-market surveillance, and the management of software changes. MHRA has also updated Recommendation 20 of the report aimed at manufacturers of AI as a medical device. It states that qualification as a medical device cannot effectively rely on wording choices or disclaimers alone and product function and purpose should be demonstrated through example outputs and user journeys. Where products sit near the qualification boundary, manufacturers are encouraged to engage with the MHRA prior to deployment.

Privacy Updates

EDPB Adopts Guidance on AI Web Scraping Anonymization and Blockchain. On July 7, 2026, the European Data Protection Board (EDPB) adopted draft guidelines on web scraping for generative AI training (legal basis for scraping; conditions for processing special category data) and on anonymization (a practical framework for assessing when data is successfully anonymized). Both guidelines are open for public consultation until October 30, 2026, and further details about the anonymization guidelines can be read in our July 2026 BioSlice Blog. The EDPB also finalized its guidelines on blockchain processing following consultation.

UK Government Opens Calls for Evidence on Data/AI Regulation and International Transfers. On July 15, 2026, the UK government launched a call for evidence on how personal and non-personal data regulation interacts with AI and other data-intensive technologies, and whether further guidance, targeted changes, or more fundamental reform is needed. A parallel call for evidence sought views on whether the UK’s international data transfer regime is enabling trusted data flows. Both close on September 9, 2026.

ICO Publishes Draft Corporate Strategy. On July 10, 2026, the Information Commissioner’s Office (ICO) published its draft corporate strategy for 2026-2028, intended to bridge its current model to the future Information Commission governance structure (transition expected in autumn 2026). Four priorities are confirmed: protecting children, promoting trust and transparency in AI, improving public sector use of personal data, and building cyber resilience. The ICO aims to set clear expectations for the responsible use of personal data in AI, to reduce regulatory uncertainty and to alleviate concerns about safety, fairness, bias, and transparency in order to realize the full benefits of AI in society. 

Product Liability Updates

UK Jurisdiction Taskforce Publishes Final Legal Statement on Liability for AI Harms. Following a public consultation, the UK Jurisdiction Taskforce has published its final legal statement on liability for AI harms under the private law of England and Wales. The statement concludes that existing English law is generally capable of addressing harms caused by AI without the need for AI-specific legislation. It identifies contract and the common law of negligence as the principal mechanisms through which liability may arise, and considers how established doctrines such as professional liability, vicarious liability, non-delegable duties, product liability, and causation apply in an AI context. The statement also addresses liability for AI-generated statements, including chatbot outputs, and confirms that outcomes will depend on established legal principles applied to the particular facts of each case. 

IP Updates

UKIPO Clarifies Patent Examination Practice for AI Inventions Following Emotional Perception. On July 14, 2026, the UK Intellectual Property Office (UKIPO) published a practice note on the search and examination of UK patent applications following the Supreme Court of the United Kingdom’s decision in Emotional Perception AI Limited v. Comptroller General of Patents, Designs and Trade Marks [2026] UKSC 3, covered in our March and May 2026 Digests. The practice note provides practical direction on how patent examiners should assess computer-implemented inventions, including inventions involving AI and artificial neural networks (ANNs). The practice notice confirms that the long-standing Aerotel approach should no longer be followed and reflects the Supreme Court’s move towards closer alignment with the European Patent Office. In particular, examiners must first determine whether a claimed invention satisfies the patent-eligibility threshold by applying an “any hardware” test. They must then identify, on a feature-by-feature basis, which aspects of the claimed invention contribute to its technical character before assessing novelty and inventive step. Only those features which contribute to the technical character of the invention may be taken into account for the purposes of inventive step. The practice note is particularly relevant for applicants seeking patent protection for AI-enabled technologies. While the Supreme Court confirmed that ANNs may constitute computer programs for the purposes of the Patents Act 1977, the UKIPO has made clear that the central question will often be whether the claimed invention makes a technical contribution capable of supporting patentability. The practice note, therefore, provides greater clarity on how AI-related patent applications will be examined in the UK, although its practical impact will depend on how the revised framework is applied during examination.