Automation bias: a growing risk in AI-assisted dentistry

UPFRONT currently lacks prospective tools to anticipate where future caries burdens will emerge. We propose that the annually updated OECD-FAO Agricultural Outlook – an instrument long used in economic planning – fills precisely this gap. Prior use of historical OECD-FAO food balance data has already demonstrated its value: Meier et al. attributed 26.3% of the global oral disease burden (4.1 million DALYs; US$172 billion in costs) to free sugars.2 The prospective projections, however, remain entirely unexploited by the oral health community. The newly released OECD-FAO Agricultural Outlook 2025–2034 reveals a widening global divide with direct implications for caries risk.3 Global sugar consumption is projected to reach 202 million tonnes by 2034, growing at 1.2% annually. Asia will account for 64% of this increase and Africa for 29% – regions where front-of-pack labelling (FoPL) and sugar-sweetened beverage (SSB) taxes remain largely absent. By contrast, consumption is already declining in Latin America and Europe, where the Outlook explicitly attributes reductions to SSB taxes and product reformulation policies.3 This divergence constitutes a natural experiment: policy-active regions are bending the sugar curve; policy-absent regions are not. By 2034, per capita sugar consumption is projected at 21.2 kg in Asia and 15.6 kg in Africa, against a global average of 23.1 kg.3 The evidence that FoPL and SSB taxes are effective – and can reduce oral health inequalities – is robust. An umbrella review of 63 studies confirmed that a 20% SSB tax reduces free sugar intake by 4.0–4.4 g/ day and childhood caries by 2.7–2.9% over ten years.4 Critically, that review identified a stark evidence gap: the impact of SSB taxation on dental caries in low- and middle-income countries – precisely the regions facing the sharpest sugar increases – has not been broadly reported.4 A 2025 Italian modelling study, covering precisely the 2025–2034 OECD–FAO forecast period, found that a 20% SSB tax would reduce DMFT by 0.07, generate €38.6 million in dental treatment savings, and produce disproportionately greater benefits in Southern Italy – a region with higher baseline caries prevalence and lower dental care utilisation – directly demonstrating equity-promoting impact.5 642 Projected per capita sugar increases of +1.2 kg in Asia and +2.6 kg in Africa by 2034,3 in the absence of proven interventions, foreshadow a substantial and preventable caries burden concentrated in the world’s least-resourced health systems. We therefore urge national dental associations, health ministries, and the WHO to: i) incorporate the annually updated OECD-FAO Sugar Chapter into routine oral health surveillance to identify emerging caries hotspots and guide resource allocation; ii) urgently commission policy-effectiveness research in Asia and Africa to address the evidence gap identified above;4 and iii) embed agricultural supply-side forecasts within global oral health monitoring frameworks as a routine caries risk outlook. The Outlook is free, methodologically transparent, and updated annually. The coming decade is the window of opportunity – dental public health must use it. B. S. Shankar, Buraydah, Saudi Arabia References 1. 2. 3. 4. 5. GBD 2021 Oral Disorders Collaborators. Trends in the global, regional, and national burden of oral conditions from 1990 to 2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet 2025; 405: 897–910. Meier T, Deumelandt P, Christen O, Stangl G I, Riedel K, Langer M. Global burden of sugar-related dental diseases in 168 countries and corresponding health care costs. J Dent Res 2017; 96: 845–854. OECD/FAO. OECD-FAO agricultural outlook 2025–2034. 2025. Available at https://www.oecd.org/en/ publications/oecd-fao-agricultural-outlook-20252034_601276cd-en/full-report.html (accessed 1 March 2026). Hajishafiee M, Kapellas K, Listl S, Pattamatta M, Gkekas A, Moynihan P. Effect of sugar-sweetened beverage taxation on sugars intake and dental caries: an umbrella review of a global perspective. BMC Public Health 2023; DOI: 10.1186/s12889-023-15884-5. Lamloum D, Dettori M, Cagetti M G, Arghittu A, Castiglia P, Campus G. Effects of a sugar-sweetened beverages tax on caries in Italy: a modelling study. Caries Res 2025; 59: 558–566. https://doi.org/10.1038/s41415-026-9893-2 Artificial intelligence Automation bias: a growing risk in AI-assisted dentistry When the output of an artificial intelligence (AI) system conflicts with our clinical judgement, do we consistently challenge the AI decision, or is there a tendency to accept it to reduce perceived risk? With the increased accuracy of AI models, the central concern is shifting. The primary risk is no longer purely technical error but a human cognitive issue, which is the gradual tendency of clinicians to accept the outputs of AI without questioning.1 This behaviour, known as automation bias or rubberstamping, can negatively affect clinical decision-making, particularly in atypical cases where human clinical judgement is essential, yet an incorrect AI output may still be accepted because the system generally performs well in routine situations and is therefore trusted blindly.2 One of the key reasons for the above issue is asymmetric risk perception. If a clinician rejects a correct AI-generated decision, they are fully responsible for that decision. In contrast, if a clinician agrees with a wrong AI output, responsibility may be seen as a system error rather than a personal one. Over time, this imbalance can lead to passive acceptance instead of active and independent critical evaluation. This highlights the need for better AI literacy among clinicians. In clinical dentistry, where AI is increasingly used for radiographic interpretation, diagnosis, and treatment planning, the risk of automation bias is highly relevant. Overreliance on algorithmic outputs could contribute to missed diagnoses and overtreatment. This emphasises the argument that AI should be positioned as augmented intelligence, supporting rather than replacing clinical judgement.3 Mitigation of automation bias requires careful design and training strategies. Explainable AI can improve transparency by clarifying how outputs are generated. It is also important to keep critical appraisal skills through education and workflow design that requires justification of decisions rather than simple acceptance. Structured exposure to incorrect AI outputs during training may also improve diagnostic vigilance. Such measures are necessary to safeguard standards of care in an increasingly AI-supported clinical environment. M. Mehrabanian, Aberdeen, UK and Liverpool, UK References 1. 2. 3. Jovchevski P, Buijsman S, Neerincx M. What is wrong with automation bias? Philos Technol 2026; DOI: 10.1007/s13347-026-01090-9. Farooqi O A. Who is really deciding? AI and clinical judgement in dentistry. Int Dent J 2026; 10.1016/j. identj.2026.109476. Mehrabanian M. Augmente (...truncated)