The performance of artificial intelligence chatbot large language models to address skeletal biology and bone health queries

Journal of Bone and Mineral Research, Journal Year: 2024, Volume and Issue: 39(2), P. 106 - 115

Published: Jan. 4, 2024

Abstract Artificial intelligence (AI) chatbots utilizing large language models (LLMs) have recently garnered significant interest due to their ability generate humanlike responses user inquiries in an interactive dialog format. While these are being increasingly utilized obtain medical information by patients, scientific and providers, trainees address biomedical questions, performance may vary from field field. The opportunities risks pose the widespread understanding of skeletal health science unknown. Here we assess 3 high-profile LLM chatbots, Chat Generative Pre-Trained Transformer (ChatGPT) 4.0, BingAI, Bard, 30 questions categories: basic translational biology, clinical practitioner management disorders, patient queries accuracy quality responses. Thirty each categories were posed, independently graded for degree four reviewers. was often able provide relevant about relevance varied widely, ChatGPT 4.0 had highest overall median score categories. Each displayed distinct limitations that included inconsistent, incomplete, or irrelevant responses, inappropriate utilization lay sources a professional context, failure take demographics context into account when providing recommendations, inability consistently identify areas uncertainty literature. Careful consideration both current AI is needed formulate guidelines best practices use as source biology.

Language: Английский

---

Desirable or Distasteful? Exploring Uncertainty in Human-Chatbot Relationships

International Journal of Human-Computer Interaction, Journal Year: 2023, Volume and Issue: 40(20), P. 6545 - 6555

Published: Sept. 12, 2023

AbstractPresent-day power users of AI-powered social chatbots encounter various uncertainties and concerns when forming relationships with these virtual agents. To provide a systematic analysis users' to complement the current West-dominated approach chatbot studies, we conducted thorough observation experienced reported in Chinese online community on chatbots. The results revealed four typical uncertainties: technical uncertainty, relational ontological sexual uncertainty. We further visibility sentiment capture response patterns toward uncertainties. discovered that identification is dynamic contextual. Our study contributes expanding, summarizing, elucidating as they form intimate AI agents.Keywords: AIsocial chatbotReplikahuman-chatbot relationshipuncertainty Disclosure statementNo potential conflict interest was by author(s).Additional informationNotes contributorsShuyi PanShuyi Pan Ph.D. student School Media Communication at Shanghai Jiao Tong University visiting researcher Utrecht University. Her research interests include chatbot, human-AI relationship, gender.Jie CuiJie Cui journalism communication & (SMC), political popular culture.Yi MouYi Mou an associate professor new media human–machine communication.

Language: Английский

---

Evaluation of the Rosa Chatbot Providing Genetic Information to Patients at Risk of Hereditary Breast and Ovarian Cancer: Qualitative Interview Study

Journal of Medical Internet Research, Journal Year: 2023, Volume and Issue: 25, P. e46571 - e46571

Published: July 20, 2023

Background Genetic testing has become an integrated part of health care for patients with breast or ovarian cancer, and the increasing demand genetic is accompanied by need easy access to reliable information patients. Therefore, we developed a chatbot app (Rosa) that able perform humanlike digital conversations about BRCA testing. Objective Before implementing this new service in daily clinical practice, wanted explore 2 aspects use: perceived utility trust technology among healthy at risk hereditary cancer how interaction regarding sensitive influences Methods Overall, 175 individuals were invited test chatbot, Rosa, before after counseling. To secure varied sample, participants recruited from all clinics Norway, selection was based on age, gender, having pathogenic variant. Among 34.9% (61/175) who consented individual interview, selected subgroup (16/61, 26%) shared their experience through in-depth interviews via video. The semistructured covered following topics: usability, usefulness, received Rosa influenced user, thoughts future use tools care. transcripts analyzed using stepwise-deductive inductive approach. Results overall finding very welcomed participants. They appreciated 24/7 availability wherever they possibility it prepare counseling repeat ask questions what had been said afterward. As created professionals, also valued as being medically correct. referred better than Google because provided specific answers questions. findings summed up 3 concepts: “Anytime, anywhere”; “In addition, not instead”; “Trustworthy true.” All (16/16) denied increased worry reading Rosa. Conclusions Our results indicate potential contribute uniform regardless geographical location. quality-assured information, tailored situation, reassuring effect our It consistent across concepts tool preparation repetition; however, none (0/16) supported could replace if confirmed. This indicates can be well-suited companion

Language: Английский

---

“I don’t see anything specifically about Black/African Americans.” Testing an Alzheimer-specific generative AI tool tailored for African American/Black communities

ACM Transactions on Computing for Healthcare, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 23, 2025

Low levels of health literacy concerning Alzheimer's Disease and related dementias (ADRD) impact African American/Black communities access to appropriate ADRD care. Additionally, a legacy mistrust in medical research due systemic racism, has resulted insufficient participation clinical trials among adults. This study explores the potential generative AI improve encourage older We designed mobile intervention featuring AI-driven conversational agents - chatbot voice assistant specifically developed for this population. tested quality using heuristics methodology adapted target population along with inputs from American/ Black professionals UX designers. Key findings highlight unique needs culturally relevant content that is accessible users varying language tailored users’ geographical location. Concerning interaction, high personalization control over interaction can promote use tool, by minimizing complexity maximizing accessibility. These show novel contribution offered our domain designing technology AI, particularly LLMS, communities.

Language: Английский

---

Integrating health equity in artificial intelligence for public health in Canada: a rapid narrative review

Frontiers in Public Health, Journal Year: 2025, Volume and Issue: 13

Published: March 18, 2025

The application of artificial intelligence (AI) in public health is rapidly evolving, offering promising advancements various settings across Canada. AI has the potential to enhance effectiveness, precision, decision-making, and scalability initiatives. However, leverage without exacerbating inequities, equity considerations must be addressed. This rapid narrative review aims synthesize related health. A methodology was used identify literature on for After conducting title/abstract full-text screening articles, consensus decision study inclusion, data extraction process proceeded using an template. Data synthesis included identification challenges opportunities strengthening 54 peer-review articles grey sources. Several applying were identified, including gaps epistemology, algorithmic bias, accessibility technologies, ethical privacy concerns, unrepresentative training datasets, lack transparency interpretability models, scaling technical skills. While advance Canada, addressing critical preventing inequities. Opportunities strengthen include implementing diverse frameworks, ensuring human oversight, advanced modeling techniques mitigate biases, fostering intersectoral collaboration equitable development, standardizing guidelines governance.

Language: Английский

---

The Digital Bridge: Using Digital Intervention to Understand Psychosocial Correlation between COVID-19 Stress and Sleep Quality in Undergraduate Students Before and After COVID-19

The Open Psychology Journal, Journal Year: 2025, Volume and Issue: 18(1)

Published: Feb. 28, 2025

Introduction With the onset of pandemic and reopening institutions, we are all undergoing a new normal, educators students attempting to adjust keep close ties core principles educational system. Existing studies have limited analysis temporal dynamics causal links between psychosocial factors, COVID-19-related stress, sleep quality. Moreover, rely on self-reported data, which introduces potential biases. Therefore, current study employs mixed-method approach that combines thematic with both inferential descriptive statistics. Methods The first part this study, is split into two phases, focuses identifying stress COVID-19 experience how it affects other behavioural, psychological, social as well sleep. It then examined significance these factors for students' academic performance during transition from offline online teaching hybrid modes. Understanding importance Digital technology and, using AI-based intervention address underlying problems, determining impact chatbots causes comprising second phase study. Results information was gathered 214 undergraduate enrolled in different programmes University Delhi self-designed, extensive questionnaire included demographic questions, Pittsburgh Sleep Index, Student Stress Questionnaire (CSSQ). To assess forecast student's based indicators, data techniques such feature selection, regression, neural networks, Naïve Bayes machine learning algorithm, multi-dimensional analysis. determine link variables before after intervention, statistical tools, including SPSS, were used calculate mean, SD, t, correlation. Conclusion results present show associated affecting sleep, their social, cognitive functioning. Additionally, research indicated significantly improved general capacity, reduced connected COVID-19,

Language: Английский

---

Can VUCA events catalyze digital public sector innovations? Evidence from three digital innovation trends in Asia

Journal of Open Innovation Technology Market and Complexity, Journal Year: 2025, Volume and Issue: unknown, P. 100529 - 100529

Published: April 1, 2025

Language: Английский

---

News Classification and Categorization with Smart Function Sentiment Analysis

International Journal of Intelligent Systems, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 24

Published: Nov. 13, 2023

Search engines are tools used to find information on the Internet. Since web has a plethora of websites, engine queries majority active sites and builds database organized according keywords utilized in search. Because this, when user types few descriptive words home page search engine, function lists websites corresponding these keywords. However, there some problems with this approach. For instance, if wants about word Jaguar, most results animals cars. This is polysemic problem that forces always provide popular but not relevant results. article presents study using sentiment technology help news classification categorization improve accuracy. We have introduced smart embedded into tackle issues record determine their sentimentality. Therefore, topic involves several aspects natural language processing (NLP) analysis for classification. A crawler was collect British Broadcasting Corporation (BBC) across Internet, carried out preprocessing text by NLP, applied methods polarity processed data. The sentimentality represents negative, positive, or neutral polarities assigned algorithms. research BBC site different explore news. toolkit (NLTK) BM25 indexed preprocessed patterns database. experimental depict proposed surpassing normal an accuracy rate 85%. Moreover, show negative Sentistrength algorithm. Furthermore, Valence Aware Dictionary sEntiment Reasoner (VADER) best-performing model obtained 85% data collected function.

Language: Английский

---

User Preferences on AI Psychotherapy Based on Moderating Effects of Individual Personality Traits: Employing a Clustering Analysis

International Journal of Human-Computer Interaction, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 18

Published: Feb. 6, 2024

The recent appearance of AI psychological counseling services is expected to lower physical, economic, and burdens for individuals seeking increase participation in the service. research still its infancy, therefore it important consider client preference achieve positive effects. However, no has been conducted investigate preferences regarding therapists. Therefore, this study explores characteristics counselors preferred by clients (e.g., rapport, trust, expertise, attraction), as well which agent therapist have effects on attitudes (users). This confirms that affect intention use counseling. Groups were divided using K-means clustering technique based individual's degree introversion/extroversion depression because desired may differ depending individual inclinations examined differences between groups. results will help enhance supplement capabilities developers actual through client-tailored

Language: Английский

---

Behaviors and Perceptions of Human-Chatbot Interactions Based on Top Active Users of a Commercial Social Chatbot

Proceedings of the ACM on Human-Computer Interaction, Journal Year: 2024, Volume and Issue: 8(CSCW2), P. 1 - 28

Published: Nov. 7, 2024

Natural language processing is enabling machines to communicate with humans naturally, yet the dynamics of extended user-chatbot interactions remain much unexplored. This study characterizes conversational styles, demographics, psychologies, and emotional tendencies most active users (i.e., top 1% by message count) a commercial chatbot platform (SimSimi.com), whom we refer as superusers. We analyzed linguistic patterns topics 1,988,971 messages written 1,994 superusers over period three years. further surveyed 76 observe their dispositions perceptions towards chatbot. find that SimSimi empathize humanize more than less users, they show higher tendency share personal negative feelings. Our findings suggest chatbots require new design considerations for who are vulnerable due high anthropomorphism openness toward machines. work also shows should have functions offer social support when necessary.

Language: Английский

---