Issue 1

Aug 3–9, 2026 · 63 stories · 6 of them from Hungary

Viewpoint

MIT study: medical AI's benefits depend on user expertise, with novices at greater risk

An MIT study finds that the benefit of medical AI assistance varies with the user's expertise, and that novices are more at risk. Another summary argues that eloquent, human-sounding AI undermines trust in transparency.

Aug 4, 2026 · 8 sources

Who is affected and how

doctor: When younger, less experienced colleagues use AI, a separate verification step should be set up, as they face the greatest risk of being misled.
hospital manager: When introducing AI support, protocols and supervision should be differentiated by training level.
health IT: Explanations in AI interfaces do not replace validation, so the presentation should be tested by user level.

regulation · European Union

EU AI Omnibus now in force: AI requirements for medical devices postponed to 2 August 2028

Regulation (EU) 2026/1744 entered into force on 27 July 2026 and postpones the high-risk requirements for medical devices containing AI to 2 August 2028. Notified bodies face a deadline of 28 January…

medtech companyThe compliance deadline for AI devices is 2 August 2028, so companies should also plan for the notified bodies' 28 January 2028 deadline in coordination with their certification body.health policyMedical AI counts as high-risk via the Annex I route, so preparation should build on the existing MDR certification structure.regulation · United States

California bill would restrict AI chatbots in mental health treatment

California lawmakers have proposed limits on the use of AI "therapists" in mental health care. Several other states are also weighing guardrails as teenagers turn to chatbots.

doctorFor psychiatrists and psychologists, it could set a legal framework for when a chatbot may be recommended as a supplement to therapy.health ITCompanies offering mental health chatbots would have to adapt their products to the proposed ban and restrictions.health policyState-level action points to a gap in federal regulation.case · United States

Consumer Reports warns AI chatbots' health answers can be inaccurate

Consumer Reports says AI chatbots give fast, understandable health information but can be inaccurate and put sensitive data at risk. A US report found that 17% of adults ask an AI chatbot health…

doctorA significant share of patients get information from AI, so at the start of a visit it is worth asking what they have read and correcting inaccuracies.health ITPrivacy risks must be addressed in patient information and in the AI features of patient portals.drug development

QuantHealth raises $45 million to run AI simulations of clinical trials

QuantHealth has raised $45 million in a Series B round. Its AI simulations let drugmakers test and adjust trial designs before enrolling patients.

pharmaceuticalsIn the trial-design phase, simulation tools can screen out designs prone to failure before patients are enrolled.

Hungary

toolCLIVORA project develops Hungarian-language AI medical speech recognition and report generationdoctorA Hungarian-language dictation and reporting system could reduce the documentation burden, provided its accuracy in Hungarian is proven.health ITHungarian-language medical speech recognition is a development direction that will have to be integrated into domestic hospital systems.caseMeskó: AI could make patients partners of their doctorsdoctorWith a better-informed patient who has prepared with AI, consultations are best run in a partnership spirit.health policyAI-supported models of patient information and data sharing can be tested in domestic pilots.toolUniversity of Szeged wins HUF 18 million grant for AI-based patient registration systemdoctorThe registration and patient-information workload at fertility centres could fall if the system reaches practice.health ITIntegrating a university-developed, AI-based registration module with clinical systems will be a separate task.clinical deploymentKertváros outpatient clinic introduces AI in patient care clinical deploymentAI is already used in imaging diagnostics at 13.2% of Polish healthcare institutionshospital managerThe regional adoption rate serves as a benchmark for evaluating AI imaging purchases.toolForestream Medical develops early-warning AI from movement patterns and clinical datadoctorThe results of clinical validation of the early-warning tool should be awaited before it is used.

In brief

clinical deploymentUpToDate Expert AI now reaches 2,500 U.S. hospitals and health systemsdoctorEvidence-based clinical decision-support AI will be widely available, so it is worth assessing its sourcing and accuracy.hospital managerAdding AI to the clinical knowledge base requires separate licensing and training decisions.insuranceMedicare AI prior authorization draws concern from patients, doctors and rights groupsdoctorWhen an AI-based denial hits a prior authorization request, document the rationale and appeal.hospital managerChanges to approval processes affect the revenue cycle and patient-pathway organization.regulationFDA's draft MDUFA VI commitment letter has implications for AI and digital healthmedtech companyCommitments tied to the user-fee cycle may affect the pace of review for AI-enabled devices, so the draft is worth reviewing.health policyThe draft shapes the regulatory framework for device authorization.regulationLondon launches AI health regulatory sandbox for NHS innovation, MHRA seeks applicantsmedtech companyAI developers can apply to the MHRA sandbox for assessment before NHS deployment.health policyThe sandbox model offers a template for bridging the gap between approval and deployment.regulationColorado law setting guardrails for AI in healthcare takes effecthospital managerColorado institutions must review their AI use practices in line with the law's requirements.health ITVendors of AI solutions must check compliance with the requirements.researchAI diagnostics research: liver cancer blood test, bowel cancer relapse risk, Parkinson's eye scans, sleep studiesdoctorThe studies point to widening screening and risk-assessment methods, and their clinical use requires further validation.researchAI breast cancer screening: DBT scores, 2D mammography risk assessment, startup grantsdoctorRadiologists need to learn the clinical interpretation of AI scores in screening and the new options for risk-based screening.medtech companyEarly-stage companies are using grant funding to move toward regulatory approval.toolIndia: AI chatbots eyed for healthcare gap, Consint.AI raises Rs 22 crore, wearable predicts heart criseshospital managerIntroducing an AI remote patient monitoring system could reduce heart failure admissions, if the reported result can be confirmed.researchGenerative biology is rewriting the rules of drug discovery, analysis sayspharmaceuticalsGenerative models are becoming part of research strategy, so skills and data should be planned ahead of time.clinical deploymentAssuta hospital network and Microsoft launch AI assistant for doctorsdoctorA hospital-network-wide AI assistant in operation shows which decision-support and documentation tasks it can take over.hospital managerA joint rollout with a major technology partner can serve as a model for structuring vendor partnerships.toolDeep Intelligent Pharma launches AI CRO 6.0 for clinical trial workflowspharmaceuticalsCRO services are being extended with AI, which affects the selection criteria for outsourced trial management.licensingNetcare submits AI clinical algorithm for approvalhospital managerA hospital group seeking approval for its own AI algorithm signals that in-house developments are also entering the regulatory pathway.toolIsraeli startup Hemispheric uses AI to diagnose brain disorders in 15 minutesdoctorFast AI-based interpretation of brain signals could complement neurological diagnostics, if clinical validation confirms it.licensingFDA actions on AI tools: breast ultrasound, intraoperative margins, hidden heart disease, Aidoc First ReaddoctorNew FDA-cleared AI tools are entering radiology and surgical practice.medtech companyThe clearance patterns offer guidance on the FDA pathway for similar AI tools.caseAI scribes in medical education: useful tool or crutch that undermines learning?doctorResidents' development of documentation skills needs separate planning when an AI scribe is used.drug developmentHassabis steps back from DeepMind to focus on Isomorphic LabspharmaceuticalsOne of the flagship AI drug-discovery ventures is getting leadership attention, which could affect partnership talks.researchScoping review examines AI-based nurse schedulinghospital managerThe review offers guidance for drawing up requirements before deployment.clinical deploymentLeeds NHS innovation and Epredia partner to spread AI pathology globally; Cambridge start-up now in 15 NHS trustsdoctorAI pathology and the assessment of scanned slides are entering the cancer diagnostic workflow.medtech companyNHS-level scaling offers an example of moving from a pilot to multi-trust deployment.drug developmentGSK to pay Relation Therapeutics up to $110 million for AI data and modelspharmaceuticalsLicensing AI models and datasets is becoming a separate procurement channel for expanding the research portfolio.drug developmentRecursion works with Genentech on first validated neuro target found through its AI mappharmaceuticalsValidation of a target found through the AI map and the scale of the partnership gauge how mature platform-based target discovery has become.drug developmentH1 and Novo Nordisk partner on AI-driven clinical developmentpharmaceuticalsThe data-platform partnership may also offer a model for AI support in investigator and site selection.regulationFDA's July non-device software update and TEMPO AI pilot draw analysismedtech companyThe device or non-device classification of AI apps needs to be reassessed in light of the FDA update.caseSurvey: 76% of Gen Z and 63% of millennials turn to AI first for health questionsdoctorYoung patients do their first round of information-seeking with AI, which reshapes expectations before the visit.health policyThe role of AI at primary care entry points must be regulated.insuranceCalifornia health insurer uses AI to boost Medicaid enrollment, but concerns remainhealth policyThe data and equity risks of AI-driven enrollment and coverage programs need separate scrutiny.insuranceOchsner Health: prior authorization automation depends on human connectionhospital managerAutomation works alongside payer relationships, so it is worth building the communication channel as well.clinical deploymentYale New Haven Health reimagines radiology for the AI eradoctorThe reshaping of radiology workflows around AI raises questions about the division of roles.toolAiforia secures €20 million EIB venture debt for AI diagnostics expansionmedtech companyThe EIB financing is an example of the capital-raising options open to European healthcare AI companies.toolOpenAI and American Psychological Association build youth mental health safeguardsdoctorInvolving a professional body indicates that chatbot protections for young people will follow professional recommendations.toolOpenAI launches ChatGPT Health globally; European access awaits regulatory clearancedoctorEuropean patients' access is delayed, but direct use may already appear through other channels.regulationIndiana HB 1271 would ban AI downcoding of dental claims; Illinois would require disclosurehospital managerIn billing and appeals, flag suspected AI-based downcoding to insurers.regulationFederation of State Medical Boards: generative medical AI should not be licensed to practice on its owndoctorAI should continue to be treated as a tool, with responsibility remaining with the physician.health policyIndividual physician licensing is not the model to follow for regulating AI.caseNHS England to revise Palantir platform metrics after FOI request reveals internal doubtshealth ITPlatform performance indicators should be measured and reported transparently.health policyFor procured digital platforms, the verifiability of performance indicators is key.clinical deploymentSAP and Fresenius partner to build AI infrastructure for healthcare in Germanyhealth ITA joint AI infrastructure from a major ERP vendor and a healthcare group bears on competition in the regional healthcare IT market.toolMiiHealth closes $2.8 million seed round; HealthSnap raises $25 millionhealth ITPatient intake and remote care AI platforms are attracting capital, which is energising the vendor market.researchAI aids bone tumour X-ray diagnosis with AUC 0.96; prostate radiation dose mapped in secondsdoctorClinical use of imaging and radiotherapy AI tools may spread after further validation.drug developmentPhylo and Chugai partner on AI agents for drug discoverypharmaceuticalsAgentic AI tools are appearing in research workflows with another major partner.clinical deploymentNGHS is an early adopter of a new AI tool for nurses regulationGhana's health ministry develops AI policy for safe use in healthcarehealth policyA national AI policy in health offers an example of how to build safety frameworks.toolOlea Medical and Median Technologies to bring AI lung cancer screening software to the US marketmedtech companyDistribution partnerships between imaging vendors and AI software makers can speed market entry.clinical deploymentAbridge and Kaiser Permanente launch Care Signals AI toolhealth ITA product built with a large US provider signals ambient AI moving beyond note-taking.

Issue 2 →

Issue 1

MIT study: medical AI's benefits depend on user expertise, with novices at greater risk

Who is affected and how

EU AI Omnibus now in force: AI requirements for medical devices postponed to 2 August 2028

California bill would restrict AI chatbots in mental health treatment

Consumer Reports warns AI chatbots' health answers can be inaccurate

QuantHealth raises $45 million to run AI simulations of clinical trials

Hungary

More key stories of the week

AI scribes raise privacy concerns, record errors and calls for a right to refuse

Survey: 90% of NHS healthcare workers use AI in clinical work

AI use for mental health support is rising among young people

AI-designed drugs advance in trials with no US approval yet; proof is the bottleneck

Nurse unions demand AI commitments in writing; New York nurses fear replacement

AI adoption outpaces governance: 63% of health systems lack an advanced AI strategy framework

Schrödinger and Bristol Myers Squibb sign deal on agentic AI drug discovery

Insilico Medicine gets China IND approval for AI-discovered Parkinson's drug candidate, launches benchmark

Stanford researchers design 16 new viruses with AI, raising biosecurity concerns

Patients use AI to read test results and arrive with self-diagnoses; doctors object

Epic Almanac: health systems move AI past pilots and start building agents, strengthening Epic's position

FDA advisory committee to review GRAIL's Galleri multi-cancer early detection test

In brief