The system is designed to address the challenges patients face between appointments, such as managing treatment side effects, addressing new symptoms, and answering complex questions about their diagnosis. By feeding insights from member responses back into the system, the partnership aims to refine future guidance and improve clinical outcomes over time. According to Amy Flaster, MD, Cigna’s chief medical officer, bringing the power of OpenAI to the group will better connect patients to specialized resources and support, whether they are dealing with the physical impact of treatment or the emotional burden of a diagnosis.

Expansion of Cigna’s Digital Health Infrastructure

The collaboration with OpenAI follows other significant artificial intelligence initiatives already established by The Cigna Group. In July, the company launched Pharmacy Forward, a $100 million AI-driven program under Accredo, designed to expand personalized care management within its health plans. The group has reported that these expanded programs have already contributed to a 42% reduction in avoidable hospital stays among engaged members. By further integrating AI in clinical workflows, Cigna seeks to continue this trend of improving healthcare efficiency through data-driven insights.

This partnership adds Cigna to a growing list of major health insurers collaborating with artificial intelligence firms to modernize their operations. Other recent industry moves include Optum’s partnership with Anthropic in July and Elevance Health’s 2023 agreement with OpenAI to provide enterprise ChatGPT access and training to its workforce. The deal reflects a broader industry shift toward utilizing generative AI to synthesize complex healthcare information and provide more timely, personalized intervention for members managing chronic or complex conditions.

Current operational bottlenecks in radiation oncology planning frequently involve coordinating complex data streams across disparate hardware and software systems. Clinical workflows often require up to 17 discrete manual stages for importing treatment parameters, and analyses of clinical cases indicate more than 44 variations in treatment planning pathways. These inefficiencies are primarily driven by unstructured information, such as clinical progress notes, pathology summaries, and disconnected imaging files. The collaboration evaluates an AI query system designed to extract and synthesize both structured electronic health record data and unstructured clinical assets, allowing clinicians to retrieve patient histories and anatomical landmarks using natural language queries.

Integration and Clinical Performance

The project evaluates a system architecture capable of parsing DICOM imaging sets, medical texts, and historical dosimetry plans into unified contextual profiles. This functional capability operates alongside the workflow orchestration engine of the iRT platform, which automates cross-department handoffs between medical physicists, dosimetrists, and radiation oncologists. By incorporating these generative AI tools, the research aims to minimize manual data curation and prevent configuration errors during the dose planning phase.

The current research builds upon previous software integration efforts between the two entities. Clinical workflow solutions implemented at Massachusetts General Hospital across more than 11,000 treatment plans successfully reduced the operational interval between patient intake and treatment initiation from 30 days to eight days. The ongoing deployment phase explores the integration of generative retrieval algorithms into broader multimodality oncology workflows, including MR-guided therapy and molecular radiotherapy guidance, to improve operational throughput in high-volume cancer centers.

Addressing Revenue Cycle Inefficiencies at the Source

Revenue cycle teams across the country continue to lose significant time and financial resources managing downstream problems — including claim denials, missing documentation, authorization failures, inaccurate charge capture, and delayed payments. Oracle Health is responding to these systemic challenges by embedding AI upstream across the full revenue cycle spectrum: from scheduling and financial clearance through clinical documentation, charge capture, billing, and payment processing.

The goal, as articulated by Oracle Health, is to identify risks early, streamline routine workflows, and connect clinical, financial, payer, and operational data before these gaps translate into payment delays and revenue loss.

Seema Verma, Executive Vice President and General Manager of Oracle Health and Life Sciences, spoke directly to the strategic intent behind the initiative: “AI gives us an opportunity to prevent revenue cycle problems before they lead to denials and delayed payments. Oracle Health is delivering a comprehensive revenue cycle solution that brings AI across the front, middle, and back office, connecting clinical and financial workflows from the first patient interaction through payment. This can help healthcare organizations reduce administrative work and revenue leakage, get paid faster, and strengthen their financial stability so they can stay focused on delivering excellent patient care.”

Connecting Fragmented Systems Through Native AI

A core challenge facing many healthcare organizations today is the fragmentation of clinical, financial, payer, provider, and operational activities across disconnected systems and manual processes. These disconnects create inefficiencies, reimbursement delays, and mounting administrative burdens.

Oracle Health’s AI revenue cycle approach addresses this by connecting workflows across patient access, clinical documentation, patient accounting, contract intelligence, payment operations, enterprise analytics, and financial management — all within existing Oracle Health infrastructure. Organizations can adopt these embedded AI capabilities at their own pace, without the disruption of wholesale system replacement.

New Embedded AI Capabilities

The newly announced AI capabilities, planned for general availability in the coming months, span five key areas of the revenue cycle:

Prior Authorization — Checks coverage, retrieves and pre-fills documentation requirements, attaches clinical evidence, coordinates authorization activities, and supports payer interactions to reduce authorization-related delays and denials.

Clinical Document Quality Integrity — Reviews reimbursement-related documentation, identifies gaps, and surfaces contextual recommendations to improve documentation and coding quality.

Charge Capture and Integrity — Analyzes clinical and reimbursement context, including modifiers and service details within charge review workflows, to limit revenue leakage and optimize reimbursement outcomes.

Medical Coding for Professional Fees — Analyzes clinical documentation and patient context to surface coding recommendations within existing workflows, supporting reimbursable claim readiness.

Appeal Management — Analyzes remittance data, identifies payment variances, and automates the creation of appeal packets when an appeal decision is made, helping reduce turnaround times.

Integration with Enterprise Financial Operations

As part of its planned product direction, Oracle Health intends to extend the connectivity of AI revenue cycle workflows into broader enterprise financial operations. This includes financial reconciliation, revenue accounting, treasury management, and analytics through platforms such as Oracle Fusion Cloud Applications. The intent is to enable health systems to connect reimbursement activity, financial operations, and enterprise analytics to support more informed operational and financial decision-making.

Strengthening Financial Stability Across Healthcare Organizations

The overarching aim of these AI revenue cycle capabilities is to help healthcare organizations reduce administrative overhead, limit revenue leakage, accelerate payments, and maintain the financial stability required to sustain quality patient care. By embedding intelligence across both the clinical and financial sides of the revenue cycle, Oracle Health is positioning its portfolio as a connected, end-to-end solution for revenue performance optimization.

The global healthcare automation tools market was valued at US$52.94 billion in 2025 and is projected to reach US$116.83 billion by 2033, representing a compound annual growth rate of 10.5% between 2026 and 2033. This expansion reflects a broader shift in healthcare technology adoption, with providers increasingly looking beyond individual automation projects toward connected systems capable of supporting multiple stages of clinical and administrative operations.

Healthcare Automation Expands Across Operational Workflows

Administrative processes remain one of the most established areas for healthcare automation. Appointment scheduling, patient intake, eligibility verification, billing, coding and claims processing involve substantial volumes of repetitive work and information exchange. Automating these activities can reduce manual intervention while allowing staff to concentrate on processes that require greater judgment and direct interaction.

The same transition is increasingly visible within clinical environments. Healthcare automation tools are being used to streamline clinical operations, coordinate information and reduce repetitive activities across different stages of care. As these technologies become connected with electronic health records and other healthcare information systems, their role is expanding from individual task execution toward broader workflow coordination.

This evolution is important because healthcare organizations rarely operate through isolated processes. A delay in one department can affect several downstream activities, while incomplete or manually transferred information can create additional administrative work. Automation that connects related workflows can therefore address operational inefficiencies at a broader level.

AI Is Moving Automation Toward Predictive Operations

Artificial intelligence is expanding the scope of healthcare automation by enabling systems to analyze operational information and identify emerging constraints.

GE HealthCare announced CareIntellect for Operations on September 15, 2026, an AI-enabled software-as-a-service application designed to provide health systems with up to a 72-hour view of emerging operational bottlenecks. The system analyzes patient and operational data, including bed availability, patient delays, staffing and wait times, and provides recommendations related to capacity and throughput.

The development reflects a broader change in the automation model. Traditional automation generally performs a predefined action after a specific trigger. AI-enabled automation can instead analyze multiple signals, identify patterns and support earlier operational intervention.

For hospitals, this distinction can be significant. Patient flow involves interconnected activities such as admissions, imaging, transfers, staffing and discharge. Automation capable of identifying relationships between these processes can potentially help operational teams respond before a bottleneck spreads across departments.

The growth of AI-enabled automation also means that future healthcare automation platforms may increasingly be evaluated according to their ability to combine information from multiple systems rather than simply automate a single task.

Laboratory Automation Is Addressing Rising Workflow Complexity

Laboratories remain an important application area because diagnostic testing involves repetitive processes that can be standardized and integrated with information systems.

Automation can extend across sample processing, testing, result management and laboratory workflow coordination. The objective is not simply to reduce manual activity but also to provide laboratories with systems capable of handling complex testing requirements and larger workloads.

Siemens Healthineers announced on September 16, 2026 that its CN-3000 and CN-6000 automated hemostasis testing systems were available for patient testing in U.S. laboratories. The platforms consolidate several hemostasis testing methodologies and are designed to automate manual tasks while connecting with track-based laboratory automation and the Atellica Data Manager for centralized oversight.

The development highlights the increasing importance of connected laboratory automation. Instead of treating analyzers as independent systems, healthcare providers can increasingly integrate instruments with workflow automation and laboratory data-management platforms.

Automation is also expanding the scale of molecular diagnostics, with automated molecular testing supporting high-throughput disease tracking and the identification of multiple genetic markers. As diagnostic laboratories manage increasingly complex testing requirements, automation can help standardize repetitive analytical processes while supporting larger testing volumes.

Imaging Automation Is Accelerating Diagnostic Workflows

Medical imaging is another area where automation and artificial intelligence are becoming increasingly interconnected.

Diagnostic imaging can generate large volumes of information that require specialist interpretation. AI-enabled tools can assist with image processing and analysis, helping clinical teams identify relevant findings and prioritize cases within established workflows.

Automated scan analysis is accelerating the processing of medical images in acute-care environments, particularly where diagnostic information needs to be assessed quickly. Automation in this area can extend beyond image interpretation to include workflow routing, image reconstruction, protocol management and reporting support.

The broader opportunity is therefore not limited to individual AI algorithms. Imaging automation can become part of an integrated diagnostic workflow in which information moves between imaging equipment, analysis software, clinical systems and healthcare professionals.

This creates an important connection between automation and interoperability. The value of an automated imaging application can depend partly on how efficiently its outputs can be incorporated into the wider clinical workflow.

Surgical Robotics Is Broadening Healthcare Automation

Robotic technology is extending automation into procedural care and creating another major area of development within the healthcare automation market.

Johnson & Johnson received FDA De Novo authorization for its OTTAVA Robotic Surgical System on July 21, 2026. The system is a table-integrated soft-tissue robotic platform authorized for multiple general-surgery procedures involving the upper abdomen. The FDA’s database records the De Novo decision, while Johnson & Johnson says the system is designed to support operating-room capacity and clinical workflow efficiency.

The development demonstrates how surgical automation is increasingly being designed around the operating environment rather than the robotic arm alone. Integrating robotics with the operating table can influence how equipment, surgical teams and physical space are organized during procedures.

AI is also becoming more closely connected with robotic surgery. Medtronic introduced Touch Surgery Aide in July 2026, an AI-enabled surgical computing platform designed to provide real-time support during procedures. The platform uses computer vision, multimodal AI and accelerated computing, while its Instrument Exit Point application provides a visual notification when selected instruments move beyond the visible field during robotic procedures.

These developments point toward a convergence of robotics, computer vision and clinical software. Surgical automation is increasingly becoming a combination of physical systems and digital intelligence rather than a purely mechanical technology.

Automation Is Moving From Individual Tools to Connected Systems

As healthcare automation expands across administrative, diagnostic and clinical environments, interoperability is becoming increasingly important.

Healthcare organizations typically operate multiple technology environments, including electronic health records, laboratory information systems, imaging platforms, medical devices and financial applications. Automation can deliver limited value when these systems remain disconnected and staff still need to manually transfer information between them.

Consequently, the next phase of market development is likely to place greater emphasis on integration. Cloud, on-premises and hybrid deployment models can support different organizational requirements, but the ability to exchange information securely and consistently across systems remains central to automation performance.

This is particularly relevant as AI becomes embedded in healthcare workflows. Automated systems need access to appropriate data, while healthcare organizations also need mechanisms for oversight, validation and governance. The integration of automation with existing infrastructure can therefore be as important as the automation capability itself.

Automation Is Extending Into Medical Device Manufacturing

The automation opportunity also extends beyond hospitals and direct care delivery into the manufacturing infrastructure supporting healthcare.

Medical-device manufacturing involves highly repetitive processes where precision, consistency and production efficiency are important. Automated tooling and assembly technologies can support the handling and positioning of components while reducing dependence on repetitive manual operations.

Precision automated tooling is helping accelerate medical-device assembly while supporting consistent handling of components, reflecting the wider expansion of automation across the healthcare technology supply chain.

This creates a broader definition of the healthcare automation market. Automation is not restricted to software used by providers; it also encompasses technologies that support the production, testing and handling of medical technologies used throughout healthcare systems.

North America Leads While Asia-Pacific Builds Momentum

North America represents the largest regional share of the healthcare automation tools market, accounting for 42.09% in the available market assessment. Its position is supported by established healthcare IT infrastructure, adoption of electronic health records and significant investment in AI, robotics and connected healthcare technologies.

Asia-Pacific accounts for 22.04% and represents the fastest-growing regional market. Healthcare digitization, expanding hospital infrastructure and increasing adoption of AI-enabled diagnostics and medical robotics are contributing to market development across the region.

Japan is an important example of automation extending across different healthcare workflows. Recent developments include AI-supported diagnostic applications and automated medication inspection technologies, illustrating how automation is being incorporated into both clinical and pharmacy-related processes.

Europe represents approximately 20% of the market, supported by established healthcare infrastructure, digital-health investment and demand for greater operational efficiency. Latin America accounts for approximately 9%, while the Middle East and Africa represent about 7%, with healthcare modernization and digital transformation creating additional opportunities for automation adoption.

The Market Is Shifting From Task Automation to Intelligent Operations

The healthcare automation tools market is increasingly defined by the convergence of several technology categories.

HHM Global observes that healthcare automation is increasingly evolving from isolated task automation toward connected systems that support clinical, diagnostic and operational workflows.

Administrative automation continues to address repetitive processes. Laboratory automation is helping manage increasingly complex diagnostic workloads. AI is extending automation into operational forecasting and clinical workflows, while robotics and computer vision are bringing automated capabilities into procedural environments.

The projected increase from US$52.94 billion in 2025 to US$116.83 billion by 2033 reflects the expansion of these applications across healthcare settings. More importantly, the direction of technology development suggests that automation is gradually becoming an underlying operational layer rather than a collection of isolated tools.

The next stage of adoption will depend on how effectively these systems can integrate with existing healthcare infrastructure while delivering measurable improvements in workflow efficiency, capacity and information management. Hospitals and other healthcare organizations are likely to increasingly assess automation according to its ability to work across departments and systems rather than its ability to automate a single repetitive activity.

As AI, robotics, diagnostics and healthcare IT continue to converge, automation is becoming embedded in more stages of the healthcare value chain. The result is a market evolving from simple task automation toward increasingly connected and intelligent healthcare operations.

The year-long programme, built around Proximie’s Intelligence Suite, targets more than 20,000 procedures annually across participating sites. The initiative aims to improve theatre efficiency and unlock additional patient capacity from existing NHS facilities, without the need for new infrastructure or extended staff working hours.

A Technology Partnership Built for Scale

The programme brings together three partners — Proximie, AWS, and Deloitte — each contributing distinct capabilities to the evaluation. Proximie leads the programme as the core technology provider. AWS is providing the cloud and AI infrastructure required to securely process high volumes of surgical data across all eight hospital sites. Deloitte is supporting programme delivery and transformation across participating NHS Trusts and the NHS London region.

Kira Levy, Head of Healthcare UK at AWS, described the initiative as “a powerful example of how cloud and AI can help hospitals to unlock capacity within existing resources, so more patients can be treated sooner.”

Sara Siegel, Global Health and Human Services Sector Leader at Deloitte, highlighted the broader context of operating theatre use: “Operating theatres are the most costly and resource-heavy part of a hospital. Efficient use of these facilities allows NHS Trusts to maximise resources and, importantly, ensure more patients receive timely procedures.”

How the Intelligence Suite Works

Proximie is a UK-headquartered health technology company that transforms operating rooms into connected ecosystems of people, devices, and data. Its cloud-based platform combines surgical telepresence, data capture, AI, and computer vision to support collaboration, training, and workflow optimisation.

The Intelligence Suite captures surgical and operational data in the background of the operating theatre without disrupting clinical teams. Data is de-identified and anonymised at source, and Proximie’s AI models then translate it into real-time insights across the operative pathway. This enables clinical teams to identify bottlenecks, reduce delays between procedures, and make better use of available theatre time.

A notable feature of the AWS infrastructure component is its ability to scale across all eight NHS hospital sites without requiring additional on-site infrastructure at each location, making the programme operationally efficient to deploy at this scale.

The Vision Behind the Evaluation

Dr. Nadine Hachach-Haram, Founder and CEO of Proximie, spoke about the motivation behind the programme following the announcement. Writing on LinkedIn, she stated: “I still operate in the NHS, and nobody in those theatres is short of effort… Sometimes there is simply a better way to do things, and it takes hospitals willing to innovate, work with us and push the NHS forward.”

She further added: “Although we are already seeing material impact from our current global deployments, this evaluation is built to find out how big an impact this makes across the NHS, at scale.”

The NHS Surgical AI pilot reflects a response to ongoing capacity challenges. The Royal College of Surgeons (RCS England) had previously expressed concern that the NHS might miss the Government’s interim target of 65% of patients starting treatment within 18 weeks by March 2026. The target was subsequently met, with 65.3% of patients starting treatment within 18 weeks in March. However, RCS England has continued to warn that sustaining that momentum remains a challenge given existing capacity constraints.

Programme Scope at a Glance

The scale of the evaluation underlines the ambition of the partnership. The programme spans 104 operating theatres, catheterisation laboratories, and endoscopy suites across 8 NHS hospital sites, runs for a duration of 12 months, targets over 20,000 procedures per year, and is led by three technology and transformation partners — Proximie, AWS, and Deloitte.

The NHS Surgical AI evaluation represents one approach being explored to make better use of existing theatre capacity, using AI and operational data to identify efficiencies across the surgical pathway. Over the next 12 months, participating sites will use insights from the Intelligence Suite to reduce procedural delays and optimise theatre scheduling, with the goal of enabling more patients to be treated without placing additional demands on clinical staff.

The partnership between Proximie, AWS, and Deloitte illustrates how cloud-based platforms and AI-driven analytics are being applied to address one of healthcare’s most persistent operational challenges — doing more with what already exists.

Known as the Manchester Sandbox, the programme provides a structured environment where innovators, healthcare professionals, and regulators can assess new technologies in real-world clinical conditions. The initiative was officially launched at MFT’s Oxford Road Campus in Manchester, bringing together representatives from the regulatory, NHS, industry, and research sectors.

The programme is expected to initially focus on AI-enabled medical devices, with the broader aim of generating robust evidence on their safety, effectiveness, and suitability for deployment within NHS care pathways. The Manchester Sandbox also intends to explore regulatory approaches that can keep pace with increasingly sophisticated medical technologies as they continue to evolve.

Faster Access to Innovation for Patients

Health Innovation Minister James Frith commented at the launch, stating: “Manchester is leading the way in getting the best new technologies into the hands of local patients faster, from AI models that can identify people at risk sooner to innovative treatments that could transform NHS care.”

The MHRA noted that the programme will enable emerging technologies to be assessed at an earlier stage, while also helping developers understand the evidence requirements needed to support their adoption within healthcare systems. For NHS healthcare technology stakeholders, including NHS organisations themselves, the sandbox is designed to offer greater insight into how new solutions perform in live operational environments and how they could realistically be integrated into existing services.

At the launch event, several technologies were highlighted as examples of the innovation the programme aims to support. These included wearable devices and AI-powered Ambient Voice Technology — the latter of which is already being used in active patient care settings.

The MHRA and MFT confirmed that learnings gathered through the Manchester Sandbox will be shared more broadly to support the safe and informed adoption of innovative technologies across the wider NHS. A roundtable is also planned, bringing together regulators, NHS organisations, innovators, researchers, and patient and public representatives to identify further opportunities for the programme.

The MHRA is expected to invite medical device developers to submit expressions of interest for the next phase of the sandbox. The partnership forms part of wider efforts to accelerate healthcare innovation while maintaining appropriate regulatory oversight, reinforcing the commitment to responsible development of NHS healthcare technology across the United Kingdom.

The National Commission into the Regulation of AI in Healthcare, established by the Medicines and Healthcare products Regulatory Agency (MHRA) in September 2025, arrived at its recommendations following an extensive evidence-gathering process involving more than 12,000 stakeholders — including clinicians, healthcare leaders, industry representatives, and technology developers.

Broad Public Support, With Clear Conditions

The commission’s findings revealed broad public support for AI in healthcare, though that support comes with clear conditions. Stakeholders consistently called for strong safety standards, meaningful human oversight, and greater transparency about when and how AI is being used in patient care. These three pillars form the backbone of the commission’s regulatory thinking.

Professor Alastair Denniston, NHS consultant and Chair of the National Commission into the Regulation of AI in Healthcare, stated: “We innovate because we believe tomorrow’s healthcare can be better than today’s. But any new innovation — including AI — needs to earn its place and earn our trust. These technologies need to show that they are safe, effective and bring benefits to patients, staff and the wider NHS, without leaving people behind.”

He added: “Over the past year we have heard from more than 12,000 people, including patients, carers, clinicians and the technologists building AI tools. One message came through clearly: people are open to AI improving their care, but only if it is safe, overseen by humans, and if they know when it is being used.”

Staged Authorisations: The ‘L-Plates’ Approach

On the matter of faster access to safe and effective AI, the commission recommends that the MHRA introduce staged authorisations for new AI models — a process it likens to ‘L-plates’ for learner drivers. Under this approach, new AI models would be deployed under close supervision with tight guardrails, allowing them to demonstrate real-world safety and performance before being granted fuller authorisation.

The commission stated that this staged approach is designed to give UK patients world-first access to promising new models, while ensuring risk is carefully controlled at every step of the process.

Continuous Real-World Monitoring Throughout the Device Lifecycle

The commission also recommends that AI-enabled medical devices be subject to continuous, real-world monitoring throughout their lifecycle, rather than evaluation at a single point in time. This recommendation reflects the evolving nature of AI technologies after they are deployed.

According to the commission, this approach would help ensure that as AI systems learn and adapt, patients remain protected and regulators retain the visibility to act quickly if problems emerge. AI-enabled medical devices present a unique regulatory challenge precisely because their behaviour can change post-deployment — making ongoing oversight a critical component of responsible AI deployment.

Public Access to Device Safety Information

Another key element of the commission’s recommendations is a provision for members of the public to easily search for information about the safety of specific AI-enabled medical devices, including any adverse incidents. The intent is to give the public “clear, accessible information” they can trust — making safety data on AI-enabled medical devices more transparent and readily available.

Strengthened Enforcement Powers for the MHRA

The commission also recommends granting the MHRA stronger enforcement powers, enabling the agency to “act decisively” in cases where AI systems “fall short of the standards patients expect.” These enhanced powers are intended to ensure that AI healthcare regulation carries meaningful authority — and that patients are protected when systems underperform or fail to meet required standards.

MHRA Chief Executive Officer Lawrence Tallon responded to the recommendations: “We have heard clearly about the opportunities of AI to speed up and improve care, and to release more of clinicians’ time for the essentially human aspects of their work with patients. We will only realise those opportunities if we have a modern, dynamic regulatory framework that accelerates safe adoption, protects patients and commands public and professional confidence. I am so grateful to the many people who contributed to this expert commission to set out the roadmap ahead.”

With the commission’s work now complete, the UK government and the MHRA will carefully consider its recommendations, with a formal response to follow in due course.

Context: Prior Policy Consultation and Current Regulatory Framework

The commission’s recommendations on safe AI rollout follow an earlier ‘policy-shaping’ consultation on AI healthcare regulation, published in June 2026. In the United Kingdom, AI in healthcare is currently regulated under the nation’s medical device regulation (MDR), with the technology classified under the designations of software-as-a-medical-device (SaMD) or AI-as-a-medical-device (AIaMD).

The consultation revealed that 50% of respondents believed the existing regulatory framework needed substantial revision, while 21% called for a complete overhaul. Key areas of concern included safety, performance standards, data governance, and privacy.

International Regulatory Developments

The MHRA is not the only regulatory body seeking industry input on AI. The United States Food and Drug Administration (FDA) also issued a call for stakeholder feedback on the use of generative AI in medical devices in August. That same month, the FDA and the MHRA announced plans for a more collaborative approach to shaping the future of healthcare regulation, with AI a key focus of that partnership — signalling growing international alignment around responsible AI deployment in clinical settings.

Why Is Hair Restoration Moving Beyond Traditional Treatments?: Conventional methods of dealing with hair loss involve the use of creams, pills, injections and, in certain cases, hair transplant surgery. These methods are still very valuable in the field of dermatology. However, some people have become wary of long-term medication, pain from injections and the risks of surgery. Furthermore, the trend is switching from “remedial” to “preventive” measures. Instead of waiting for bald spots to appear, people are taking care of their scalp and hair roots. And this creates a niche for technologies that are supposed to facilitate the body’s processes of hair growth.

What Makes Red Light Therapy Different?: The way red light therapy works is by means of photobiomodulation and the use of low-level red or near-infrared light that stimulates biological activity, but does not cause harm or heat the tissue. So, it is different from hair restoration surgery that involves follicle transplant and incision making. What makes it appealing is that it is non-invasive and easy to carry out. Studies have been conducted into its application in androgenetic alopecia or pattern hair loss. Clinical research proves the increase of hair density in some patients, though results might differ and evidence can vary for different types of alopecia.

How Does It Work at a Cellular Level?: Photobiomodulation is theorized to act at the cellular level. Specific wavelengths of red and near-infrared light are thought to interact with mitochondria the energy-producing organelles within cells and specifically with an enzyme called cytochrome c oxidase, which plays a role in mitochondrial respiration. Some research suggests this interaction may encourage hair follicles to enter or remain in the growth phase, delaying the transition to the resting (telogen) phase. That said, the exact mechanism is still not fully understood, and this remains an active area of study.

Why Is Technology-Enabled Hair Wellness Gaining Attention?: The rising fascination with red light therapy points to a wider trend in the sphere of wellness. People are now getting used to the idea of using technology when maintaining their health on a daily basis. Hair care also falls into this trend. At-home devices that use light support the idea that red light therapy can be very convenient. This does not mean that this treatment is effective, for there are many important aspects that can influence the results of the therapy, such as wavelength and duration of the therapy.

Could Red Light Therapy Shape the Future of Hair Restoration?: Looking ahead, experts anticipate less emphasis on finding entirely new alternatives and more focus on combining proven treatments with emerging ones like photobiomodulation. Studies are being conducted about the most favorable wavelengths, schedules of treatment, delivery methods and the effectiveness of photobiomodulation along with other methods. Now, red light therapy can be considered a very promising non-invasive technology in terms of hair loss treatment, especially in the case of patterned alopecia, but not as a universal remedy. Looking ahead, experts anticipate less emphasis on finding entirely new alternatives and more focus on combining proven treatments with emerging ones like photobiomodulation.

To conclude, red light therapy stands for an interesting change in the procedure of hair restoration because the emphasis is laid on the health of follicles and cell activity. While research on the therapy continues in order to prove its long-term efficacy, the non-invasive character along with the enhancement of scientific support for it make the therapy very promising in modern dermatology. The future of hair restoration will probably be more personalized and technology-oriented.

By Sonam Kumari

The next challenge for healthcare AI is not simply generating clinical notes. It is generating notes that clinicians can trust.

Generative AI and ambient AI scribes are increasingly being explored to reduce the administrative burden of clinical documentation. By converting clinician-patient conversations into structured notes, these systems could give clinicians more time to focus on patient care.

But a well-written note is not necessarily an accurate one.

An AI system can leave out a medication, change the meaning of a statement, confuse a past condition with a current one, or introduce information that was never discussed. In healthcare, these errors can affect patient safety, workflow, compliance, and accountability.

My analysis of 1,965 dialogue-note examples from MTS-Dialog, ACI-Bench, and PriMock57 illustrates why clinical AI needs to be evaluated beyond traditional language-generation measures. Factual accuracy, omissions, hallucinations, clinical concept retention, editing effort, evidence, and human review all matter.

Why Traditional AI Metrics Are Not Enough

Metrics such as ROUGE can measure how closely an AI-generated note resembles a reference note. However, similarity in wording does not necessarily mean that the clinical meaning is correct.

Consider a simple example. If a patient says, “I have no chest pain,” an AI system could generate, “Patient reports chest pain.” The sentence is fluent, but the clinical meaning has been reversed.

Similar errors can occur when a system omits a medication dosage, changes the certainty of a diagnosis, attributes a patient statement to the clinician, or leaves out follow-up instructions.

The benchmark results reinforce this concern. In the MTS-Dialog analysis, the strongest configuration achieved a ROUGE-1 score of 42.52 compared with 30.42 for the BART-large baseline. However, the same configuration had a factual F1 score of 0.7675, a 3% hallucination rate, and a 33% omission rate.

The implication is important: improving language quality does not automatically make clinical documentation safer.

Designing a More Reliable Workflow

A reliable clinical documentation system should be designed as an end-to-end workflow rather than simply connecting an audio recording to a large language model.

The process can be organized into five stages.

Capture and transcription. The system first converts the clinical conversation into text while preserving speaker identity, sequence, timestamps, and confidence information. Errors at this stage can affect everything that follows.

Clinical fact extraction. Before generating a narrative note, the system should identify information such as symptoms, diagnoses, medications, allergies, laboratory results, findings, temporal information, negations, and treatment plans. These facts should remain connected to the original conversation whenever possible.

Section-aware generation. The system can then generate different sections of the clinical note, including the history of present illness, objective findings, assessment, and plan. Each section has different accuracy requirements. For example, the history needs to preserve symptom chronology and important negative findings, while the plan needs to accurately capture medications, referrals, investigations, and follow-up instructions.

Evidence and provenance. Important statements should be traceable to the conversation that supports them. If the system generates information about a medication or treatment plan, clinicians should be able to identify its source.

Clinician verification. The generated note should remain a draft until it is reviewed and approved by an appropriately qualified clinician. The clinician should be able to identify missing information, correct errors, review supporting evidence, and approve the final record.

This approach positions AI as a documentation assistant rather than an autonomous author of the medical record.

Hallucinations Are Not the Only Problem

Healthcare AI discussions often focus on hallucinations, but omissions deserve similar attention.

In the analysis, one configuration achieved factual precision of 0.9408 with a hallucination rate of only 1%. However, its omission rate was 37%. Another configuration reduced omission to 33%, but its hallucination rate increased to 3%.

This illustrates an important trade-off. A system can avoid unsupported information while still producing an incomplete record. Conversely, a system that attempts to capture more information may increase the risk of introducing unsupported statements.

Clinical documentation therefore needs a balanced evaluation framework. Factual precision and recall should be considered alongside hallucination rates, critical omissions, and retention of clinically important concepts.

Human Review Should Be Part of the Product

Human oversight is sometimes viewed as a limitation of AI. For clinical documentation, it should instead be treated as a product requirement.

AI can help identify information that deserves additional attention, including medication names and dosages, allergies, diagnoses, abnormal findings, negations, follow-up instructions, and low-confidence transcription segments.

This can make review more targeted. Instead of asking clinicians to scrutinize every sentence equally, the system can direct attention toward information where an error could have greater consequences.

The goal is not to eliminate clinical accountability. It is to reduce administrative work while preserving the clinician’s responsibility for the final record.

What Healthcare Organizations Should Measure

Organizations evaluating clinical documentation AI should look beyond model benchmark scores.

Useful measures include documentation time saved per encounter, clinician editing time, acceptance rates, unsupported claims per note, critical errors, medication and allergy omissions, negation and temporal errors, clinician satisfaction, transcription failure rates, system availability, and model-version tracking.

These measures help answer a more important question: does the technology actually improve the clinical workflow?

A model can perform well on a benchmark while creating additional review work for clinicians. Deployment success is ultimately a workflow outcome, not simply a model benchmark outcome.

Privacy and Governance From the Start

Clinical conversations contain sensitive information, so privacy and governance should be considered from the beginning of product development.

Production systems should incorporate appropriate access controls, encryption, audit logging, retention policies, recording consent, model-version tracking, provenance, and controlled integration with electronic health records.

Testing should also reflect the diversity of real clinical environments. Accents, languages, speech conditions, and different patient populations can affect system performance. Results from a controlled English-language dataset may not predict how a system performs in practice.

Moving From Benchmarks to Real-World Validation

Benchmark datasets provide valuable information for comparing systems, but they cannot capture every condition found in clinical encounters.

Real consultations can involve interruptions, background noise, overlapping conversations, accents, incomplete histories, specialty-specific terminology, and rapidly changing clinical situations. The datasets used in this analysis also have limitations in size and representativeness.

Future evaluation should therefore include real clinical environments, multiple specialties, diverse patient populations, different languages, EHR integrations, clinician correction time, clinically significant errors, privacy considerations, and provenance accuracy.

Recent research is already examining how clinicians modify AI-generated drafts in real-world settings, reinforcing the importance of studying not only what AI generates but also what clinicians need to correct.¹

Five Questions for Healthcare AI Leaders

Healthcare organizations evaluating clinical documentation AI should consider five questions:

1. Can the system show where important information came from?
2. Does it measure omissions as well as hallucinations?
3. Can clinicians review and correct the output efficiently?
4. Are workflow and safety outcomes measured alongside model performance?
5. Are privacy, governance, and auditability built into the product?

These questions shift the discussion from whether an AI model can generate a note to whether the complete system can be trusted in practice.

Conclusion

The next phase of clinical AI will not be defined simply by how well machines can write clinical notes. It will depend on whether those notes can be verified, corrected, monitored, and trusted.

Generative AI has significant potential to reduce the documentation burden on clinicians. However, healthcare requires more than fluent text. AI systems need to preserve clinical meaning, recognize uncertainty, provide evidence for important statements, and keep clinicians responsible for the final record.

The most valuable clinical AI may not be the system that tries to replace the clinician’s documentation process. It may be the one that handles repetitive work while making important information easier for clinicians to review.

The real opportunity is not simply generating more clinical notes. It is using AI to help create documentation that is accurate, explainable, and safe.

About the Author

Sonam Kumari is a technology and product professional with a background in computer science engineering, data analytics, healthcare technology, AI-enabled product development, and enterprise platforms. Her work focuses on applying artificial intelligence, data, and product strategy to complex healthcare and enterprise technology challenges.

Alongside the Epic integration, OpenAI also released its Healthcare Public Data plugin, a dedicated tool that directs ChatGPT to draw answers from nine official healthcare datasets, including PubMed, DailyMed, and the CMS Coverage Database.

Connecting Clinical Data Across Systems

In its official announcement, OpenAI outlined the core challenge the partnership seeks to address. “Healthcare organizations need AI that works across the systems and information central to care and operations,” OpenAI stated. “Patient context, medical evidence, public healthcare data, and organizational knowledge often live in different places. Connecting these sources in a governed workspace helps teams find the right information, understand it in context, and put it to work across the business.”

Through the ChatGPT EHR integration, clinicians can now query patient-specific information directly within the AI environment. Use cases include identifying which screenings a patient is due for or determining which lab results a clinician should review ahead of an appointment. In select deployments, ChatGPT will be embedded directly into the EHR workflow, allowing providers to access AI capabilities without leaving the patient chart.

UCSF Health Among First Pilot Partners

UCSF Health served as a pilot partner on the project, giving the integration a real-world clinical testing ground. Suresh Gunasekaran, president and CEO of UCSF Health, commented on the technology’s potential in the press release: “By bringing relevant information together more quickly and comprehensively, the technology has the potential to reduce time spent synthesizing data and give clinicians more time with patients. We’re also engaging frontline teams to validate these capabilities in practice and help shape where they can add the most value.”

Healthcare Public Data Plugin: Nine Trusted Sources

The Healthcare Public Data plugin represents a focused expansion of ChatGPT’s capabilities within clinical and research environments. Rather than relying solely on general training data, the plugin directs the AI to pull responses from nine verified public health resources. These sources are ClinicalTrials.gov, the CMS Coverage Database, CMS Open Data, DailyMed, openFDA, Medicare Care Compare, the National Plan and Provider Enumeration System, PubMed, and RxNorm.

OpenAI described the plugin as building on ChatGPT’s existing ability to help teams answer clinical questions and synthesize medical research. The OpenAI healthcare offering now goes further by anchoring responses in official, publicly available data. A research team, for instance, could use ClinicalTrials.gov to identify trials in active recruitment, while a pharmacist could consult DailyMed for the most recent medication warnings. Teams performing multidisciplinary work, such as population health teams, can query all nine sources simultaneously.

Safety and Accuracy Evaluated Across Clinical Use Cases

OpenAI emphasized the safety and accuracy standards behind the ChatGPT EHR integration. The company noted that it worked alongside physician partners to evaluate the connected EHR context across 27 clinical use cases. Across a total of 4,363 ratings, physicians rated 99.1% of ChatGPT responses as safe across all use cases. Additionally, 93% of tested responses were rated as having “good” or better accuracy.

The dual release of the Epic Systems integration and the Healthcare Public Data plugin positions ChatGPT as a more structured and verifiable resource for clinical teams, supported by real-world pilot data and direct integration into established healthcare workflows.