Current and Potential Applications of Ambient Artificial Intelligence

Abstract

The fourth industrial revolution has transformed our daily life, with the introduction of new digital technologies.1Xu M. David J.M. Kim S.H. The fourth industrial revolution: opportunities and challenges.Int J Financ Res. 2018; 9: 90-95Crossref Scopus (451) Google Scholar This revolution has led to the integration of digital and physical worlds, created value, and impacted every sector of the economy. Ambient artificial intelligence (AI) is one such technology that is enabled by the fourth industrial revolution and has a great future potential for augmenting health care delivery.2Mahmood M.R. Raja R. Kaur H. Kumar S. Nagwanshi K.K. Ambient Intelligence and Internet of Things: Convergent Technologies. 1st ed. Scrivener Publishing, 2022Crossref Scopus (0) Google Scholar The need for digital interventions and augmentation in health care delivery and clinical medicine has never been greater. The postpandemic, current endemic society has crippled health care institutions around the world. Physician burnout and resignations are at an all-time high.3Bhardwaj A. COVID-19 pandemic and physician burnout: ramifications for healthcare workforce in the United States.J Healthc Leadersh. 2022; 14: 91-97Crossref PubMed Scopus (8) Google Scholar Physician shortages in the United States and around the world have also peaked. Ancillary clinical staff are suffering from the same conditions. Digital technologies in clinical medicine and in health care delivery are needed now more than ever. Cook et al,4Cook D.J. Augusto J.C. Jakkula V.R. Ambient intelligence: technologies, applications, and opportunities.Pervasive Mob Comput. 2009; 5: 277-298Crossref Scopus (755) Google Scholar described ambient intelligence as a presence of digital environment that is sensitive, adaptive, and responsive to the presence of people. Ambient intelligence integrates human-centric computer interfaces, secure systems and devices, and technologies that assist with sensing, reasoning, and acting.4Cook D.J. Augusto J.C. Jakkula V.R. Ambient intelligence: technologies, applications, and opportunities.Pervasive Mob Comput. 2009; 5: 277-298Crossref Scopus (755) Google Scholar As illustrated in Figure 1, ambient intelligence works at the intersection of Internet of Things devices and sensors placed in the user’s surrounding environment, pervasive computing, AI, machine learning, and human-computer interaction. In addition to machine learning, ambient intelligence also uses knowledge graph–based technologies.5de Vargas M.F. Pereira C.E. Ontological User Modeling for Ambient Assisted Living Service Personalization. Springer, 2017Crossref Scopus (1) Google Scholar Ambient sensing exists as the result of the use of various ambient sensors, such as video cameras, depth, thermal, radio, acoustic, and wearable sensors (eg, in smart watch). The perception data from the ambient sensors, when integrated with various AI solutions, can help to design early warning systems to prevent adverse safety events, improve and support decision making, provide clinical workflow and operational efficiencies, and relieve administrative burden for physicians. Ambient AI applications in health care delivery, as shown in Figure 2, can be leveraged at various touch points, such as outpatient clinics, hospital, and patient’s home/daily living space to augment health care delivery, by benefiting patients and their direct care providers/family members, clinical care providers, and enterprises. Ambient clinical intelligence: This is a conversational AI application that integrates ambient voice sensing technology and virtual assistant function, to automate and streamline the visit documentation into the electronic health record (EHR), and data retrieval from the EHR by a physician, during a patient’s clinic encounter.6Nahar J.K. Lopez-Jimenez F. Utilizing conversational artificial intelligence, voice, and phonocardiography analytics in heart failure care.Heart Fail Clin. 2022; 18: 311-323https://doi.org/10.1016/j.hfc.2021.11.006Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar This application can be used at various point-of-care settings, including outpatient offices, and can also be integrated with virtual visit (Telemedicine) platforms. The touchless voice activated virtual assistant helps to decrease the administrative burden of physicians, enables better physician-patient interaction, and improves patient satisfaction and experience. Ambient AI technology is being used in health care to create virtual nurse assistants that can provide personalized care to patients. These virtual assistants can be integrated into patients’ homes, hospitals, and clinics, providing real-time assistance to patients, monitoring their health status, and communicating with health care providers. For example, in a hospital setting, virtual nurse assistants can monitor patients’ vital signs and alert the health care providers if there are any changes in their condition. They can also provide patients with information about their medications, assist with scheduling appointments, and even remind patients to take their medication. In a home setting, virtual nurse assistants can provide patients with reminders to take their medication and follow their treatment plans. They can also monitor patients’ health status and alert the health care providers if there are any concerning changes. One example of a virtual nurse assistant is Sensely’s “Molly,” a conversational AI platform that provides patients with personalized health coaching and care management. Patients can interact with Molly through a smartphone application or a smart speaker, and the platform uses natural language processing to understand patients’ needs and provide personalized support. There are various benefits of ambient AI–enabled virtual assistants. The virtual nursing assistants can lower health care costs by reducing the need of in-person nursing staff visits and preventing hospital readmissions by timely triage and enabling care coordination with the physicians.7Lee D. Yoon S.N. Application of artificial intelligence-based technologies in the healthcare industry: opportunities and challenges.Int J Environ Res Public Health. 2021; 18: 271Crossref PubMed Scopus (147) Google Scholar Physician facing virtual assistants in the form of digital scribe can help in decreasing physician burnout.8Wang J. Lavender M. Hoque E. Brophy P. Kautz H. A patient-centered digital scribe for automatic medical documentation.JAMIA Open. 2021; 4: ooab003https://doi.org/10.1093/jamiaopen/ooab003Crossref PubMed Scopus (8) Google Scholar Overall, ambient AI is transforming the health care sector by providing new ways to deliver personalized care to patients, and administrative assistance to physicians. There are two important hospital spaces where ambient AI has great utility. These spaces are intensive care or critical care unit (ICU) and operating rooms. i.First is to prevent data fatigue and augment the workflow of clinicians. Mayo clinic built an EHR interface for clinicians in the ICU called Ambient Warning and Response Evaluation. This ambient intelligence–based application enabled filtering of meaningful data from the vast volume of data entering into the EHR, delivering real-time context specific, high value clinical information to the physicians, augmenting timely clinical decision support, and preventing data overload.9Herasevich V. Pickering B. Gajic O. How Mayo Clinic is combating information overload in critical care units. Harvard Business Review.https://hbr.org/2018/03/how-mayo-clinic-is-combating-information-overload-in-critical-care-units#:∼:text=A%20rules%2Dbased%2C%20ambient%2D,insights%20from%20clinicians%20and%20patientsDate accessed: December 30, 2022Google Scholarii.Second is monitoring of patient mobilization. Ambient sensors installed in ICU rooms can help in evaluating patient movements, detect use of external assistance, and interactions with physical space such as sitting on a bedside chair. Ambient intelligence can be used in future to study the relationship between patient mobilization, length of stay, and patient recovery.10Haque A. Milstein A. Fei-Fei L. Illuminating the dark spaces of healthcare with ambient intelligence.Nature. 2020; 585: 193-202Crossref PubMed Scopus (116) Google Scholariii.Third is infection control. To prevent hospital acquired infections, ambient sensors can help to monitor hand washing activities.10Haque A. Milstein A. Fei-Fei L. Illuminating the dark spaces of healthcare with ambient intelligence.Nature. 2020; 585: 193-202Crossref PubMed Scopus (116) Google Scholar Chen et al,11Chen J. Cremer J.F. Zarei K. Segre A.M. Polgreen P.M. Using computer vision and depth sensing to measure healthcare worker-patient contacts and personal protective equipment adherence within hospital rooms.Open Forum Infect Dis. 2015; 3ofv200Crossref PubMed Scopus (8) Google Scholar have studied the use of computer vision and depth sensing sensors to measure health care worker-patient contacts and personal protective equipment adherence within hospital rooms. The authors concluded that using the computer vision and depth sensing, we can estimate potential hand hygiene opportunities at the bedside and estimate adherence to personal protective equipment. Insights gained from research studies using Ambient intelligence can guide changes in behavior of hospital staff, lead to better infection control strategies in hospital units, lower morbidity rates, and reduce length of stay, resulting in better patient outcomes. Use of ambient cameras acquired video during surgery, and computer vision has potential in evaluation of surgical skills of the surgeons, facilitate timely feedback for refining surgical skills, which can improve technical efficiency, and decrease the complication rate.12Khalid S. Goldenberg M. Grantcharov T. Taati B. Rudzicz F. Evaluation of deep learning models for identifying surgical actions and measuring performance.JAMA Network Open. 2020; 3e201664Crossref Scopus (65) Google Scholar Another potential use of ambient AI is the use of ambient cameras to automate surgical tool count, for preventing accidental retention of surgical instruments inside the patient and its associated complications. In daily living spaces, ambient AI has 4 potential uses as mentioned further.i.Ambient assisted monitoring and fall prevention: ambient sensors in living spaces for the older population can help to monitor daily living activities, impairment of which can be associated with increase in risk of fall, with its adverse health consequences.13Luo Z, Hsieh J-T, Balachandar N, et al. Computer vision-based descriptive analytics of seniors’ daily activities for long-term health monitoring. Machine Learning for Healthcare (MLHC). 2018;2(1)Google Scholar Ambient intelligence systems with integrated contactless sensors in daily living spaces of older populations can help in prompt detection of fall, provide alerts to the caregivers, facilitate emergency response, and provide timely intervention to prevent morbidity and mortality.ii.Ambient assisted rehabilitation and disease surveillance: ambient sensors can be used for gait analysis, which can help in designing optimum home rehabilitation programs for patients recovering from stroke. Ambient sensors can also be used at home, for patients with Parkinson disease. Yang et al,14Yang Y. Yuan Y. Zhang G. et al.Artificial intelligence-enabled detection and assessment of Parkinson’s disease using nocturnal breathing signals.Nat Medicine. 2022; 28: 2207-2215Crossref PubMed Scopus (44) Google Scholar studied the use of AI-based system that integrates contactless ambient radio sensor for detecting Parkinson's disease, predicting disease severity, and tracking disease progression over time using nocturnal breathing. Authors in this study found that AI system using ambient sensor can identify people who have Parkinson's disease from their nocturnal breathing and can accurately assess their disease severity and progression.iii.Ambient assisted living: Assisted living technologies based on ambient intelligence are called ambient assisted living tools, which can be used for improving quality of life for elderly and people with functional diversity.5de Vargas M.F. Pereira C.E. Ontological User Modeling for Ambient Assisted Living Service Personalization. Springer, 2017Crossref Scopus (1) Google Scholar These may include voice assistants and smart home solutions that can help with automated task performance for people with motor disability, timely medication reminders, and provide adaptations for patients with cognitive impairment. Social robots can help to address loneliness among the older population by providing social company and promoting social engagement with their relatives.iv.Ambient assisted working: with an increase in aging workforce that is challenged by physical and cognitive limitations and chronic health disorders, there is an emerging role of Ambient Intelligence–Ambient Assisted Working. These ambient assisted working tools and solutions can use wearable and environmental sensors, cyber-physical systems, and AI to provide flexible workplace adaptations and offer support and adaptation to older working population in a variety of workplace scenarios, helping them to fulfill different work-related activities.15Spoladore D. Trombetta A. Ambient assisted working solutions for the ageing workforce: a literature review.Electronics. 2023; 12: 101Crossref Google Scholar Ambient AI applications in health care delivery can create value by promoting quadruple aim, which includes improved clinician experience, improved patient experience, lower cost, and better outcome, as shown in Figure 3. As we look forward to designing, developing, and scaling ambient AI applications in health care delivery, the following challenges need to be considered and optimally addressed as applicable to the use case, keeping human-centered AI, and multistakeholder perspectives in mind. Various ambient sensors may capture a large volume and variety of personal/private data, which the patient may not be comfortable in sharing. There is risk of inappropriate sharing of data without the patient’s knowledge and use of data beyond its original intended use or by third parties without informed consent. Ambient AI applications used in clinical practice should be fair and bias free. The potential for bias is a recognized challenge for the implementation of AI applications in health care.16Martinez-Martin N. Luo Z. Kaushal A. et al.Ethical issues in using ambient intelligence in health-care settings.Lancet Digit Health. 2021; 3: e115-e123https://doi.org/10.1016/S2589-7500(20)30275-2Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar, 17Char D.S. Shah N.H. Magnus D. Implementing machine learning in health care—addressing ethical challenges.N Engl J Med. 2018; 378: 981Crossref PubMed Scopus (576) Google Scholar, 18Challen R. Denny J. Pitt M. Gompels L. Edwards T. Tsaneva-Atanasova K. Artificial intelligence, bias and clinical safety.BMJ Qual Saf. 2019; 28: 231-237Crossref PubMed Scopus (367) Google Scholar Ambient AI applications may also be subjected to these biases, including bias in data sets (input data) and algorithmic bias. Beyond the computational and statistical biases, 2 important categories of bias that are overlooked are human bias and systemic bias. Human bias includes cognitive bias that affects decision making of the individuals involved in designing and developing the ambient AI solutions. The systemic bias is operational at the level of the entire health care institution/enterprise, which has practices or norms that result in the favoring or disadvantaging of certain social groups.19Schwartz R. Vassilev A. Greene K. et al.Towards a standard for identifying and managing bias in artificial intelligence.NIST Spec Publ. 2022; 1270: 1-77Google Scholar Attention should be given to addressing bias and ensuring fairness, while designing, developing, and deploying the ambient AI applications. For patients and physicians to trust ambient AI applications, it is important to ensure data and algorithmic transparency. There should be clarity regarding data set composition, mode of data collection, and process of annotation and how the data will be used and shared. Model outputs should be interpretable and clinically relevant. An important challenge is ensuring explainability, which implies the ability of the ambient AI system to clearly explain to the end users its prediction and decision-making process.20Markus A.F. Kors J.A. Rijnbeek P.R. The role of explainability in creating trustworthy artificial intelligence for health care: a comprehensive survey of the terminology, design choices, and evaluation strategies.J Biomed Inform. 2021; 113103655Crossref PubMed Scopus (203) Google Scholar In context of explainability it is important to mention about Explainable AI (XAI), which is a research field that aims to make AI systems results more understandable to human.21Adadi A. Berrada M. Peeking inside the black-box: a survey on explainable artificial intelligence (XAI).IEEE Access. 2018; 6: 52138-52160Crossref Scopus (2449) Google Scholar Explainability and transparency, help in developing trustworthy ambient AI applications for physicians and patients. Ambient AI applications should not be unethically used causing harm or adversity to the patients. Important stakeholders should be included in the design and development phase to ensure ethical implementation. Ambient AI governance framework should include ethical use as an essential component.22Reddy S. Allan S. Coghlan S. Cooper P. A governance model for the application of AI in health care.J Am Med Inform Assoc. 2020; 27: 491-497https://doi.org/10.1093/jamia/ocz192Crossref PubMed Scopus (200) Google Scholar Increase in number of ambient sensors and the volume of data associated with continuous ambient sensing introduces legal challenges, such as who is responsible when there is delay in response or when there is lack of timely response to the Ambient AI system?23Gerke S. Yeung S. Cohen I.G. Ethical and legal aspects of ambient intelligence in hospitals.JAMA. 2020; 323: 601-602https://doi.org/10.1001/jama.2019.21699Crossref PubMed Scopus (30) Google Scholar Additionally, who bears the liability if there is an adverse clinical outcome due to an error in the output of the ambient AI system? Patients may not be comfortable with use of ambient AI applications due to concern about loss of privacy given the thought of being continuously monitored.16Martinez-Martin N. Luo Z. Kaushal A. et al.Ethical issues in using ambient intelligence in health-care settings.Lancet Digit Health. 2021; 3: e115-e123https://doi.org/10.1016/S2589-7500(20)30275-2Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar There may be additional concerns about security and unauthorized sharing of personal and sensitive data, which may be used to their disadvantage. Physicians may not be comfortable using these systems, due to lack of familiarity, technical knowledge, and trust. During design and development of the ambient AI applications, lack of human-centered approach, especially in older population who may have physical, mental limitations, and lack of comfort with technology use, can be a challenge for successful adoption.24Auernhammer J. Human-centered AI: the role of Human-centered Design Research in the development of AI. Paper presented at: Synergy—DRS International Conference; August 11-14, 2020; Brisbane, Australia.Google Scholar Looking in the should be given to the of human-centered the ambient AI application to the needs of the end and the context of use, will help in providing and J. Human-centered AI: the role of Human-centered Design Research in the development of AI. Paper presented at: Synergy—DRS International Conference; August 11-14, 2020; Brisbane, Australia.Google Scholar This will help to address the challenges of the human-centered in of smart with use of ambient AI applications to improve patient in outpatient clinics, increase of physicians and by decreasing administrative data and prevent adverse patient by using early warning A. M. smart health to smart and 1st ed. Springer, Scopus Google of physicians and important stakeholders the health care to clinical of ambient AI applications and sharing of This may help to address the end among physicians in these of data development of optimum governance privacy and informed with regarding data use and sharing. This will help address the challenges of fairness, transparency, and ethical use of ambient AI N. Luo Z. Kaushal A. et al.Ethical issues in using ambient intelligence in health-care settings.Lancet Digit Health. 2021; 3: e115-e123https://doi.org/10.1016/S2589-7500(20)30275-2Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar Ambient AI is a emerging which is in the early phase of has great potential to over the The use of ambient sensing technology and intelligence in the of will provide to a patient’s health care the health care delivery of and to practice intelligence-based

Ähnliche Arbeiten