Development and validation of the Dental Artificial Intelligence Readiness and Acceptance Instrument (DAI-RAI) for dental professionals Authors: Narmin Helal, Mohammed Ghazi Aljohani, Ahmed Ghazi Aljohani, Osama Adel Basri, Ola B. Al-Batayneh, Bahn Agha, Maryam Quritum, Mohanid Almozughi, Mohammad Zeinalddin, Nader Abdulhameed, Hanaa Mohammed Alhalki, Heba Jafar Sabbagh Published: 2026 ABSTRACT BackgroundArtificial intelligence (AI), including applications such as radiographic image analysis, caries detection, and treatment planning, is increasingly integrated into dental diagnostics, education, and clinical workflows. However, validated instruments to evaluate dental professionals’ preparedness for AI remain limited, particularly those capturing both readiness and acceptance in clinical contexts. This study developed and psychometrically validated the Dental Artificial Intelligence Readiness and Acceptance Instrument (DAI-RAI) to assess readiness and acceptance toward AI adoption in dentistry.MethodsItems were constructed based on the Technology Readiness Index (TRI) and Technology Acceptance Model (TAM), adapted to the dental AI context, and refined through expert review and pilot testing. The final validation sample included 941 dental professionals. Exploratory factor analysis (EFA) using Maximum Likelihood with Varimax rotation examined factor structure, followed by confirmatory factor analysis (CFA) using AMOS and R. Internal consistency and convergent validity were evaluated using Cronbach’s alpha (α), Composite Reliability (CR), and Average Variance Extracted (AVE).ResultsSampling adequacy was confirmed using the Kaiser–Meyer–Olkin (KMO) measure, which was excellent (KMO = 0.95; Bartlett’s χ²= 6640.32, p < 0.001). EFA supported a two-factor solution accounting for 61.8% of the variance, and one low-loading item (loading = 0.37) was removed based on the predefined threshold of ≥ 0.40. CFA indicated an acceptable overall fit (CFI = 0.915; TLI = 0.900; SRMR = 0.047), although RMSEA (0.108) suggested moderate model misfit. The final validation was conducted on a large sample of dental professionals (N = 941). The final instrument, comprising two modified constructs, AI-Technology Readiness (AI-TR) and AI-Technology Acceptance (AI-TA), consisted of 15 items (AI-TR = 8, AI-TA = 7). The DAI-RAI demonstrated excellent reliability and convergent validity (AI-TR: α = 0.92, AVE = 0.621; AI-TA: α = 0.94, AVE = 0.659). Partial measurement invariance was established across professional roles.ConclusionThe DAI-RAI is a concise, reliable, and theory-grounded measure that evaluates dental professionals’ AI readiness and acceptance. Its validated structure supports its application in educational planning, workforce development, and the ethical implementation of AI in dental care.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12903-026-08556-4. Background Artificial intelligence (AI), including machine and deep learning, is increasingly transforming dental practice, with applications in diagnostics, treatment planning, and clinical workflows [1, 2]. AI-powered tools can detect caries and periodontal disease, assist in implant and orthodontic planning, and manage patient records. These advancements improve diagnostic accuracy and clinical efficiency while optimizing outcomes and resource use [3]. However, AI integration presents challenges, including ethical considerations like bias and transparency, data privacy and security concerns, and the need to upskill dental professionals [4]. Addressing these obstacles is crucial for the responsible and equitable implementation of AI in dental practice. Given the transformative impact of AI in dentistry, it is essential to address the factors that influence its effective integration, particularly dental professionals’ readiness, knowledge, and adoption of these technologies. Successful integration of AI in dental practice affects its acceptance, utilization, and overall impact [5]. Successful AI integration depends on professionals’ readiness, knowledge, and adoption behavior, which together influence how technologies are understood, accepted, and implemented in practice [6, 7]. These factors determine the rate and quality of AI adoption; higher levels of preparedness and knowledge enable quicker, more efficient adoption, leading to better results [8]. A lack of preparedness, understanding, or knowledge may hamper AI’s potential in dentistry. To understand and predict the adoption of AI in dentistry, it helps to consider established theoretical frameworks such as the Technology Readiness Index (TRI) and the Technology Acceptance Model (TAM). The TRI assesses an individual’s propensity to adopt new technologies through four dimensions: optimism, innovativeness, discomfort, and insecurity, the latter reflecting concerns about potential negative consequences [9]. TAM, on the other hand, focuses on how users accept and use technology. Its core constructs are: perceived usefulness, the degree to which a person believes that using a particular technology would enhance their job performance; perceived ease of use, the degree to which a person believes that using a particular technology would be effortless; and behavioral intention to use, the intention to perform a specific behavior, such as using the technology. TAM posits that perceived usefulness and ease of use influence behavioral intention, which, in turn, leads to actual use. Both models have been applied to healthcare technologies like electronic health records and telehealth [10]. They can be adapted for dental applications to explore factors influencing AI acceptance, given the field’s unique features [11]. TRI helps identify individual differences to inform targeted interventions, while TAM emphasizes how perceptions of usefulness and ease of use affect adoption [10]. Despite the use of these models, gaps still exist in assessing AI readiness in dentistry. Existing tools, such as the TRI and the TAM [10], provide insights into technology adoption but fall short in fully assessing AI readiness among dental professionals. Specifically, these models often lack AI-specific assessment, such as evaluating concerns related to algorithm transparency, data privacy, model bias, and the need for specialized training in AI-driven clinical tools [4, 12], and insufficiently consider the specific context of dental practice [5]. This includes factors like diverse treatments, the dentist-patient relationship, and regulatory issues. To address these limitations, there is a need for a context-specific instrument that integrates established models, TRI and TAM, with domain-relevant considerations. The proposed Dental Artificial Intelligence Readiness and Acceptance Instrument (DAI-RAI) provides a structured framework to assess AI readiness and acceptance, while capturing practical barriers and contextual factors relevant to dental practice [1]. The instrument evaluates dental professionals’ preparedness, perceptions, and understanding of AI, and identifies perceived benefits, implementation challenges, and ethical or regulatory considerations. Although the focus is on dental professionals, the DAI-RAI may also be applicable to other groups within dentistry, such as undergraduate students, who are increasingly exposed to digital technologies and AI-driven learning environments. This integration provides a structured, validated foundation to advance educational, clinical, and policy initiatives that promote safe, ethical, and patient-centered AI adoption in dentistry. Materials and methods Study design and overview This study employed a multi-phase, cross-sectional design to develop and psychometrically validate the DAI-RAI. The development process was conducted in four sequential stages: (1) item generation (2), expert evaluation for content validity (3), pilot testing, and (4) main validation involving exploratory and confirmatory factor analyses. Each phase was designed to ensure theoretical rigor, content relevance, and statistical robustness in measuring dental professionals’ readiness and acceptance of AI technologies. Figure 1 illustrates the four-phase development and validation process of the DAI-RAI, including item generation, expert content validation, pilot testing, and main psychometric validation with corresponding sample sizes at each stage. Fig. 1Development and validation of the dental artificial intelligence readiness and acceptance instrument (DAR-RAI) for dental professionals Development and validation of the dental artificial intelligence readiness and acceptance instrument (DAR-RAI) for dental professionals Ethical approval for this study was obtained from the Research Ethics Committee of the Faculty of Dentistry, King Abdulaziz University, Jeddah, Saudi Arabia (REC No: 57-05-25). All participants were informed of the study’s objectives and provided electronic informed consent prior to participation. Data were collected anonymously, and all procedures adhered to the principles outlined in the Declaration of Helsinki. Phase 1: Item generation Items were derived from established literature on the TRI and TAM, as well as on the application of AI in dentistry [13]. The items were selected for their relevance to AI applications in healthcare and dentistry and their alignment with theoretical frameworks. The TRI and TAM frameworks were integrated to capture both readiness and acceptance of AI. TRI domains (optimism, innovativeness, discomfort, and insecurity) reflect individuals’ predisposition toward technology [14]. Optimism and innovativeness captured confidence and engagement with AI, whereas discomfort and insecurity reflected concerns related to anxiety, ethics, and data privacy. Meanwhile, TAM constructs (perceived usefulness, perceived ease of use, and behavioral intention) assess acceptance and intended use in clinical practice [4]. Behavioral intention to use AI was also included as a central TAM domain, reflecting the likelihood that positive perceptions translate into future implementation and sustained clinical use. These adapted dimensions, informed by prior studies [15, 16], ensured that the DAI-RAI comprehensively captured both readiness and acceptance toward AI within the dental context. An initial pool of 30 items was drafted, encompassing these theoretical and contextual domains. Each item was phrased in plain, practice-relevant language to ensure clarity and inclusivity across experience levels. These items were checked for face validity by two independent researchers, who reviewed all items for clarity, relevance, and appropriateness of wording. Also, consensus between both researchers was required for retention. Phase 2: Expert evaluation (content validation) Three academic consultants, two in pediatric dentistry and one in orthodontics, evaluated the preliminary instrument for clarity, relevance, representativeness, and comprehensiveness. Content validity was established through expert review. They independently rated each item for relevance on a 4-point scale. The Item-Level Content Validity Index (I-CVI) was calculated as the proportion of experts rating each item as 3 or 4. Items with I-CVI < 0.78 were revised or removed. The Scale-Level Content Validity Index (S-CVI/Ave) was 0.92, indicating excellent overall content validity. Their collective clinical and academic expertise supported a critical refinement process: redundant or ambiguous items were removed, while AI-specific statements emphasizing ethical implications and data protection were added. The instrument was refined from 30 to 22 items prior to pilot testing. The provisional instrument was organized into two established theoretical frameworks, TRI and TAM, ensuring that AI-specific constructs were conceptually embedded within a robust adoption model. Phase 3: Pilot testing A pilot test was conducted with 18 dental professionals, selected via convenience sampling to ensure representation across training and experience levels, including students, interns, general dentists, and specialists. Participants evaluated linguistic clarity, interpretability, and response consistency by completing the survey online. Minor wording adjustments were made to enhance flow, while no items were removed. The finalized pre-validation version was then prepared for full psychometric testing. Phase 4: Main validation procedure Participants and data collection The DAI-RAI instrument was administered electronically in English via Google Forms. A total of 941 dental professionals participated in the validation phase. The sample included dental students (35.7%), interns, general dentists, and specialists from Saudi Arabia and neighboring Arab countries. The mean age was 29.5 ± 9.0 years, and approximately 66% were female. Most participants (63%) worked in academic or government institutions, ensuring representation of individuals with digital literacy and exposure to AI technologies. Surveys were administered electronically; informed consent was obtained before participation. Statistical analysis All analyses were conducted using SPSS version 29.0 (IBM Corp., Armonk, NY, USA) and AMOS version 29.0, with confirmatory procedures verified in R (lavaan package, version 0.6–17). Exploratory Factor Analysis (EFA) EFA was performed using Maximum Likelihood extraction and Varimax rotation. Factors were retained based on eigenvalues > 1, scree-plot inspection, and item loadings ≥ 0.40 with communalities ≥ 0.30. Sampling adequacy was excellent (Kaiser–Meyer–Olkin (KMO) = 0.95), and Bartlett’s test confirmed factorability (χ² = 6640.32, p < 0.001). Reliability and convergent validity Internal consistency was evaluated through Cronbach’s α, average inter-item correlations, and Composite Reliability (CR). Average Variance Extracted (AVE) was computed to assess convergent validity, adopting the criterion AVE ≥ 0.50. Confirmatory Factor Analysis (CFA) The factor solution derived from the EFA was tested using Maximum Likelihood estimation. Model fit was assessed via Comparative Fit Index (CFI), Tuker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR), with theoretical justification guiding any model refinement. Measurement invariance Multi-group CFA evaluated invariance across gender, educational level, and professional role using sequential configural, metric, and scalar models. Invariance was confirmed when ΔCFI ≤ 0.01 and ΔRMSEA ≤ 0.015. All analyses adhered to STROBE guidelines, with statistical significance set at p < 0.05. The development and validation of the DAI-RAI followed COSMIN standards for health measurement instruments. Results Sample characteristics The validation sample comprised 941 dental professionals representing a wide range of clinical and academic roles. The mean age was 29.5 ± 9.0 years, with approximately two-thirds female. Participants included dental students (35.7%), interns, general dentists, and specialists from Saudi Arabia (n = 254, 27%), Jordan (n = 200, 21.3%), Iraq (n = 162, 17.2%), Egypt (n = 111, 11.8%), Libya (n = 115, 12.2%), and others (n = 99, 10.5%). Most respondents (63%) were affiliated with academic or government institutions, ensuring that the instrument was evaluated among professionals exposed to digital and AI-based technologies in dentistry (Table 1). These characteristics ensured adequate representation across early- and mid-career dental professionals, supporting generalizability of the validation results. Table 1Demographic and professional characteristics of the validation sample (n = 941)CharacteristicCategoryn ( % ) or Mean ± SDAge (years)—29.5 ± 9.0GenderFemale626 ( 66.5 )Male315 ( 33.5 )Professional levelDental students336 ( 35.7 )Interns162 ( 17.2 )General dentists263 ( 27.9 )Specialists180 ( 19.2 )Institutional affiliationAcademic or public sector593 ( 63.0 )Private sector174 ( 18.4 )Other174 ( 18.6 )Summary of the validation sample (n = 941) Table 1Demographic and professional characteristics of the validation sample (n = 941)CharacteristicCategoryn ( % ) or Mean ± SDAge (years)—29.5 ± 9.0GenderFemale626 ( 66.5 )Male315 ( 33.5 )Professional levelDental students336 ( 35.7 )Interns162 ( 17.2 )General dentists263 ( 27.9 )Specialists180 ( 19.2 )Institutional affiliationAcademic or public sector593 ( 63.0 )Private sector174 ( 18.4 )Other174 ( 18.6 )Summary of the validation sample (n = 941) Demographic and professional characteristics of the validation sample (n = 941) Summary of the validation sample (n = 941) Exploratory Factor Analysis (EFA) Sampling adequacy was excellent (KMO = 0.95), and Bartlett’s test confirmed factorability (χ² = 6640.32, df = 105, p < 0.001). Using Maximum Likelihood extraction with Varimax rotation, a clear two-factor solution emerged, corresponding to AI-Technology Readiness (AI-TR) and AI-Technology Acceptance (AI-TA). Items loaded ≥ 0.57 with acceptable communalities. One item was relocated to the AI-TR construct based on its higher factor loading on the readiness factor and its conceptual alignment with innovativeness and preparedness, as determined by pattern matrix inspection during EFA. One negatively worded AI-TR item (“I feel apprehensive about using new AI technologies in dentistry”) showed a low loading (0.37) and a weak item–total correlation (r = 0.14) and was removed prior to CFA. The final EFA retained 15 items (AI-TR = 8 items; AI-TA = 7 items), collectively explaining 61.8% of the total variance across the two factors (Table 2). Table 2Exploratory Factor Analysis (EFA) diagnostics for the DAI-RAI (n ≈ 420)StatisticValueKMO0.95Bartlett’s test of sphericityχ² = 6640.32 (df = 105, p < 0.001*)Extraction methodMaximum LikelihoodRotation methodVarimaxFactors retained2 (Technology Readiness and Technology Acceptance)Total variance explained61.8%**Items removedLow loading (0.37) and r = 0.141 itemExploratory factor analysis supported a two-factor solution consistent with Technology Readiness and Technology Acceptance. Sampling adequacy and sphericity were excellentKMO Kaiser–Meyer–Olkin* Significant: factorable correlation matrix** Acceptable construct coverage. One low-performing item (TR4) was removed before CFA (TR4: “I feel apprehensive about using new AI technologies in dentistry”) Table 2Exploratory Factor Analysis (EFA) diagnostics for the DAI-RAI (n ≈ 420)StatisticValueKMO0.95Bartlett’s test of sphericityχ² = 6640.32 (df = 105, p < 0.001*)Extraction methodMaximum LikelihoodRotation methodVarimaxFactors retained2 (Technology Readiness and Technology Acceptance)Total variance explained61.8%**Items removedLow loading (0.37) and r = 0.141 itemExploratory factor analysis supported a two-factor solution consistent with Technology Readiness and Technology Acceptance. Sampling adequacy and sphericity were excellentKMO Kaiser–Meyer–Olkin* Significant: factorable correlation matrix** Acceptable construct coverage. One low-performing item (TR4) was removed before CFA (TR4: “I feel apprehensive about using new AI technologies in dentistry”) Exploratory Factor Analysis (EFA) diagnostics for the DAI-RAI (n ≈ 420) Exploratory factor analysis supported a two-factor solution consistent with Technology Readiness and Technology Acceptance. Sampling adequacy and sphericity were excellent KMO Kaiser–Meyer–Olkin * Significant: factorable correlation matrix ** Acceptable construct coverage. One low-performing item (TR4) was removed before CFA (TR4: “I feel apprehensive about using new AI technologies in dentistry”) Reliability Internal consistency was excellent for AI-TA (α = 0.94) and very good for AI-TR (α = 0.89). After removing the low-performing AI-TR item, reliability increased to α ≈ 0.92, confirming a stable construct (Table 3). These findings supported proceeding with confirmatory factor analysis on the independent subsample. Table 3Reliability indices of the DAI-RAI subscalesSubscaleNo. of itemsCronbach’s αα after refinementComposite reliability (CR)Average variance extracted (AVE)AI-TR80.890.920.9120.621*AI-TM70.94—0.9380.659**Cronbach’s alpha values and composite reliability indicate strong internal consistency for both subscales* Very good internal consistency** Excellent internal consistency. Removing the low-performing AI-TR4 item improved the readiness subscale reliability Table 3Reliability indices of the DAI-RAI subscalesSubscaleNo. of itemsCronbach’s αα after refinementComposite reliability (CR)Average variance extracted (AVE)AI-TR80.890.920.9120.621*AI-TM70.94—0.9380.659**Cronbach’s alpha values and composite reliability indicate strong internal consistency for both subscales* Very good internal consistency** Excellent internal consistency. Removing the low-performing AI-TR4 item improved the readiness subscale reliability Reliability indices of the DAI-RAI subscales Cronbach’s alpha values and composite reliability indicate strong internal consistency for both subscales * Very good internal consistency ** Excellent internal consistency. Removing the low-performing AI-TR4 item improved the readiness subscale reliability CFA The entire sample (N = 941) was randomly divided a priori into two independent subsamples. These were an EFA subsample (n = 420) and a CFA subsample (n = 521). EFA was performed only on the first subsample, and CFA on the second, to prevent overfitting and preserve cross-validation integrity. The CFA subsample (n = 521) showed an acceptable fit for the two-factor model (AI-TR + AI-TA): χ² = 627.06, df = 89, p < 0.001; CFI = 0.915; TLI = 0.900; RMSEA = 0.108 (90% CI 0.100–0.116); SRMR = 0.047. Although the CFI and SRMR indicated acceptable model fit, the RMSEA exceeded the 0.08 threshold. This suggests some degree of model misfit and should be interpreted with caution. Standardized loadings were strong for AI-TA (0.78–0.91) and robust for most AI-TR items (0.77–0.89). One AI-TR item (AI-TR2) had moderate loading (0.57) and high uniqueness (u² = 0.67). However, it was retained because it captures a theoretically important facet of readiness (preparedness amid uncertainty) that would otherwise be absent from the construct. (Table 4). Overall, the CFA presented partial support for the two-factor structure underlying the DAI-RAI; however, future studies may further refine the model. Table 4Confirmatory Factor Analysis (CFA) model fit indices (n = 521)Fit indexValueRecommended thresholdχ² (df)627.06 (89) p < 0.001*—Comparative Fit Index (CFI)0.915≥ 0.90Tucker–Lewis Index (TLI)0.900≥ 0.90Root Mean Square Error of Approximation (RMSEA)0.108 (90% CI 0.100–0.116)≤ 0.10Standardized Root Mean Square Residual (SRMR)0.047≤ 0.08* Confirmatory factor analysis showed an acceptable two-factor model fit for AI-Technology Readiness and AI-Technology Acceptance using maximum-likelihood estimation Table 4Confirmatory Factor Analysis (CFA) model fit indices (n = 521)Fit indexValueRecommended thresholdχ² (df)627.06 (89) p < 0.001*—Comparative Fit Index (CFI)0.915≥ 0.90Tucker–Lewis Index (TLI)0.900≥ 0.90Root Mean Square Error of Approximation (RMSEA)0.108 (90% CI 0.100–0.116)≤ 0.10Standardized Root Mean Square Residual (SRMR)0.047≤ 0.08* Confirmatory factor analysis showed an acceptable two-factor model fit for AI-Technology Readiness and AI-Technology Acceptance using maximum-likelihood estimation Confirmatory Factor Analysis (CFA) model fit indices (n = 521) * Confirmatory factor analysis showed an acceptable two-factor model fit for AI-Technology Readiness and AI-Technology Acceptance using maximum-likelihood estimation Multi-group confirmatory factor analysis was conducted on the CFA subsample (n = 521) to test invariance across professional subgroups (e.g., students, general dentists, specialists). The configural and metric models demonstrated stable structure and loadings (ΔCFI ≤ 0.01, ΔRMSEA ≤ 0.015), supporting full metric invariance across groups (Table 5). These findings suggest that the DAI-RAI measures readiness and acceptance equivalently across varying professional levels. Table 5Measurement invariance across professional subgroupsModel constraintCFIRMSEAΔCFIΔRMSEAConfigural0.9000.117——Metric0.9010.1120.0010.005Scalar0.8990.1100.0030.002Strict0.8940.1120.0040.002Multi-group CFA showed configural and metric invariance, indicating consistent factor structure and loadings across subgroupsCFI Comparative Fit Index, RMSEA Root Mean Square Error of Approximation Table 5Measurement invariance across professional subgroupsModel constraintCFIRMSEAΔCFIΔRMSEAConfigural0.9000.117——Metric0.9010.1120.0010.005Scalar0.8990.1100.0030.002Strict0.8940.1120.0040.002Multi-group CFA showed configural and metric invariance, indicating consistent factor structure and loadings across subgroupsCFI Comparative Fit Index, RMSEA Root Mean Square Error of Approximation Measurement invariance across professional subgroups Multi-group CFA showed configural and metric invariance, indicating consistent factor structure and loadings across subgroups CFI Comparative Fit Index, RMSEA Root Mean Square Error of Approximation Final instrument structure The validated DAI-RAI retained a two-factor structure (AI-TR and AI-TA) comprising 15 items after refinement. AI-TR captured optimism, innovativeness, and preparedness toward AI adoption, whereas AI-TA assessed perceived usefulness, ease of use, and intention to integrate AI tools (Supplementary Tables S1 and S2, Supplementary File). Discussion This study developed and validated the DAI-RAI as a structured instrument to assess readiness and acceptance of artificial intelligence in dentistry. The findings support a coherent two-factor framework capturing both predispositional readiness and system-specific acceptance, reflecting distinct yet complementary components of technology adoption. Two item-level decisions warrant explicit discussion. First, we retained one AI-TR item with a moderate loading (0.54) because it captures a theoretically important facet of readiness, preparedness despite uncertainty. Retaining one sub-0.60 item is justifiable in early validation when (a) scale reliability stays high (TRI α ≈ 0.92), (b) convergent validity is acceptable (AVE > 0.50), and (c) the item provides conceptual coverage that would otherwise be lacking. This approach is consistent with recommendations in psychometric practice and with evidence that TRI’s facets can contribute differentially to usage via mediated pathways [14, 17]. Second, we removed a negatively worded readiness item (“AI in dentistry makes me nervous”) because it showed weak communality and lowered item-total correlation, an expected pattern for reverse-scored items in professional samples, where wording effects can mimic factors and decrease internal consistency. Eliminating that item improved reliability and interpretability without altering the theorized two-factor structure, which remained intact in the confirmatory model. The two-factor solution identified in this study aligns clearly with established models of technology adoption: AI-TR captures individuals’ likelihood to adopt emerging technologies, while AI-TA models beliefs about a system’s perceived usefulness and ease of use. Our framework, therefore, reflects the well-known pathway from general readiness to system-specific acceptance. Although TRI is traditionally modeled with four dimensions, factor loadings in this study supported a unidimensional readiness construct, consistent with recommendations that collapsing readiness facets can improve parsimony in early clinical adoption settings. This matches the findings of the large meta-analysis by Blut & Wang [14], which showed that TRI’s four dimensions collectively explain technology usage and influence downstream acceptance factors, supporting TRI as a readiness predictor for adoption models. Similarly, Ammenwerth’s [16] health-informatics review confirms that TAM constructs (usefulness, ease of use) have consistently predicted the adoption of clinical information systems, reinforcing our use of TAM as the acceptance component in a healthcare setting. Recent empirical research in dentistry has also found that dentists’ technology readiness independently predicts the extent of their adoption of digital tools in practice, underscoring the need for targeted readiness efforts to accelerate digital integration [18]. Although their study focused on general digital technologies rather than AI specifically, the findings are consistent with our view that readiness and acceptance are distinct but complementary factors in adoption. Likewise, Alotaibi et al. [19] underscored the significance of TAM determinants, perceived usefulness, and ease of use, in adopting digital dental technologies and dental informatics, reinforcing our TAM framework within a dental setting. Additionally, the AI literature in dental education highlights that knowledge gaps and training deficits are common, which explains why readiness and acceptance alone do not automatically translate into AI use without adequate infrastructure and curricula [17, 20–22]. This study identifies adoption determinants that other scales overlook. A large cross-sectional study demonstrated that technology readiness independently predicts the extent of digital integration in practice, highlighting it as a necessary upstream condition for adoption, yet did not model AI-specific attitudes or barriers. Complementing this, the study done by Alotaibi et al. [19] reported that classic TAM determinants, perceived usefulness and perceived ease of use, shape dentists’ intentions to adopt digital dental technologies and informatics, again at a general (non-AI-specific) level. By explicitly operationalizing AI-focused constructs, the DAI-RAI advances existing technology-adoption frameworks in several ways: (i) it separates baseline predisposition to adopt technology from system-specific evaluations within a unified measurement structure; (ii) it contextualizes these constructs within AI-driven clinical environments, addressing the unique cognitive and ethical aspects of AI adoption, such as trust, perceived reliability, and data transparency, highlighted in recent healthcare adaptation studies and (iii) it offers a validated, dentistry-specific instrument capable of assessing both readiness and motivational intention to use AI tools, thereby filling a key gap in current digital-health measurement literature. The need for this AI-specific lens is well supported by recent reviews in dentistry and health informatics. Evidence syntheses show that, while usefulness and ease of use predict uptake for conventional digital systems, AI introduces additional frictions, notably data privacy and legal constraints, model bias and transparency concerns, and the requirement for new competencies, making mere transposition of generic digital-adoption tools insufficient [12, 23]. In dental education, the study conducted by Thurzo et al. [22] documents persistent knowledge and training gaps that limit AI readiness despite positive attitudes toward digital technology. Thus, compared with prior work that treats “digital technology” as a single category, our AI-specific scale targets the higher technical complexity, ethical sensitivity, and resource burden unique to AI systems, considerations that are increasingly cited as decisive for real-world adoption in dentistry. The DAI-RAI is a single, context-specific instrument grounded in two integrated constructs, AI-TR and AI-TA. These modified applications of the TAM and TRI models in health informatics capture how readiness and acceptance influence adoption in dental practice. By aligning readiness with intention, the DAI-RAI helps explain why positive attitudes toward emerging technologies do not always lead to actual use, a gap often observed in digital health research. This alignment strengthens predictive validity and provides a clearer view of how professional motivation develops from initial acceptance to consistent application in clinical and educational contexts. Practically, the instrument supports multiple levels of implementation. Academic institutions can use it to assess students’ and educators’ preparedness, ensuring that digital learning is introduced effectively. At the organizational level, it enables benchmarking of readiness and acceptance across departments or regions. For policymakers and curriculum developers, findings from the DAI-RAI can guide the development of evidence-based strategies for the ethical and patient-centered integration of AI technologies, encouraging responsible innovation and trust in clinical environments [3]. Overall, the DAI-RAI moves beyond general digital-readiness surveys by offering a validated framework that connects perceptions, preparedness, and behavioral intention, helping educators, researchers, and healthcare leaders translate acceptance into meaningful, sustainable adoption. The DAI-RAI aligns with findings from similar adaptations of the TRI and TAM frameworks in other healthcare fields, such as medicine, radiology, and nursing informatics, where readiness and perceived usefulness consistently predict technology adoption [10, 16]. However, while these studies often focus on general digital health tools, the DAI-RAI uniquely captures the nuances of emerging AI systems, reflecting the specific combination of trust, ethical confidence, and clinical adaptability needed for integration into dentistry. Consistent with prior research in medical imaging, professionals show optimism about the benefits of technology but remain cautious about autonomy and accountability, emphasizing the importance of addressing the “Insecurity” and “Discomfort” dimensions during implementation [12]. Compared with prior single-country investigations, the multinational sampling in the current study enhances generalizability and provides a broader understanding of readiness across diverse professional and cultural contexts. This diversity offers valuable insight into the differing levels of technological exposure and regulatory preparedness across regions. In Middle Eastern contexts, where AI deployment in healthcare remains in its early stages, challenges related to ethical governance, workforce training, and regulatory awareness are particularly relevant [19]. The DAI-RAI captures these realities by providing a standardized framework that reflects both global trends and region-specific readiness profiles. The validated dimensions of the DAI-RAI hold practical value for educators and practitioners seeking to foster digital competence within dentistry. For educators, the results highlight areas where AI-related learning outcomes and experiential training can be incorporated into undergraduate and postgraduate curricula. Assessing students’ readiness and perceived ease of use using the AI-TR and AI-TA subscales can guide the development of tailored teaching interventions that improve confidence and reduce discomfort with AI-driven digital innovations. Clinicians and practice owners can also use the instrument to assess their teams’ readiness before adopting new AI systems, enabling more strategic investments and targeted upskilling initiatives. Of particular importance are the “Insecurity” and “Discomfort” dimensions, which reflect underlying concerns about data governance, professional accountability, and loss of clinical control. Addressing these through structured training, transparent communication, and evidence-based protocols can enhance both competence and trust. Building readiness, therefore, is not merely a technical goal; it is an educational and cultural process that depends on leadership, ethical awareness, and collective confidence in technology’s role in improving care quality and patient safety. Limitations and future directions This study has several limitations that should be acknowledged. Its cross-sectional and self-reported design may introduce bias and limit causal interpretation. While the DAI-RAI demonstrated robust psychometric performance across a large and diverse sample, cultural and linguistic variations could have influenced item interpretation. Although most model fit indices indicated an acceptable fit, the slightly elevated RMSEA (0.108) suggests that refining or testing higher-order constructs may further improve model fit. Future work should strengthen the instrument’s evidence base through longitudinal validation and test–retest reliability to assess temporal stability. Cross-cultural invariance testing across countries and practice settings would help verify measurement equivalence and ensure broader generalizability. Applying the DAI-RAI within intervention studies could evaluate how readiness and acceptance evolve following structured AI education or implementation programs. Finally, incorporating behavioral indicators, such as real-world AI technology utilization and performance outcomes, would enhance predictive validity and bridge the current gap between intention and actual adoption. Conclusion The DAI-RAI is a validated, theory-based instrument that enhances understanding of how dental professionals perceive and prepare for AI digital transformation. It offers a structured and reliable way to evaluate readiness, acceptance, and confidence in adopting AI emerging technologies. Beyond its psychometric strength, the DAI-RAI provides a foundation for education, policy, and clinical practice, supporting the ethical and evidence-based integration of technology while maintaining patient safety and professional integrity. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3.