Biases in human perception of facial age are present and more exaggerated in current AI technology

Ganel, Tzvi; Sofer, Carmel; Goodale, Melvyn A.

doi:10.1038/s41598-022-27009-w

Cited by 4 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, although artificial intelligence (AI) technology can deliver comprehensive characteristics of facial aging, it has not been widely used owing to considerable errors and biases. 31 , 32 The use of questionnaire-based tools can help overcome the challenges associated with facial aging assessment. 33 The large sample size of the IEU OpenGWAS project may alleviate concerns about measurement errors when evaluating facial skin aging using questionnaires, highlighting the benefits of such data.…”

Section: Discussionmentioning

confidence: 99%

Genetically Proxied Autoimmune Diseases and the Risk of Facial Aging

Zhang,

Li,

Geng

et al. 2024

CCID

View full text Add to dashboard Cite

Purpose Previous studies have suggested a relationship between autoimmune diseases and the risk of facial skin aging. However, evidence from population-based studies on this topic is limited, leaving the causal association between these factors unknown. This study aimed to systematically evaluate the causal effects of 18 autoimmune diseases on the risk of facial skin aging, aim of providing strategies to mitigate early facial aging in patients with autoimmune diseases. Patients and Methods We conducted univariate Mendelian randomization (UVMR) analyses to examine the causal relationship between 18 autoimmune diseases and facial aging using publicly available summary data from genome-wide association studies (GWASs). We also conducted multivariate Mendelian randomization (MVMR) analyses to adjust for confounding factors, including smoking, alcohol consumption, and body mass index (BMI). Results The main inverse variance weighted (IVW) method revealed that genetically proxied ankylosing spondylitis (AS) (OR 1.017; 95% CI: 1.003–1.031; P =0.018), sicca syndrome (SS) (OR 1.008; 95% CI: 1.005–1.011; P = 2.66×10 −6 ), systemic lupus erythematosus (SLE) (OR 1.006; 95% 1.001–1.011; P =0.014), multiple sclerosis (MS) (OR 1.004; 95% CI: 1.001–1.007; P =0.021), primary sclerosing cholangitis (PSC) (OR 1.002; 95% CI: 1.000–1.004; P =0.023), and celiac disease (CeD) (OR 1.002; 95% CI: 1.001–1.004; P =0.009) were significantly associated with higher risk of facial aging. After adjusting for potential confounding factors, the association persisted between AS, SLE, and CeD. Conclusion These findings indicated that autoimmune diseases play a causal role in facial skin aging. Therefore, patients with autoimmune diseases should take appropriate measures to prevent early facial aging.

show abstract

Section: Discussionmentioning

confidence: 99%

Genetically Proxied Autoimmune Diseases and the Risk of Facial Aging

Zhang,

Li,

Geng

et al. 2024

CCID

View full text Add to dashboard Cite

show abstract

“…First, many of these types of bias (racism, sexism, and ageism) have their roots in similar problems, such as underrepresentation [4,9]. For example, a 2022 study of 21 age-recognition systems found that artificial intelligence systems consistently identified age with less accuracy across all age, gender, and ethnic categories, which the authors speculated was due to older adults being underrepresented in training data [49]. Second, the methods used to balance age-related bias in these papers may also have applications for other types of bias.…”

Section: Overviewmentioning

confidence: 99%

Strategies to Mitigate Age-Related Bias in Machine Learning: Scoping Review

Chu,

Donato-Woodger,

Khan

et al. 2024

JMIR Aging

View full text Add to dashboard Cite

Background Research suggests that digital ageism, that is, age-related bias, is present in the development and deployment of machine learning (ML) models. Despite the recognition of the importance of this problem, there is a lack of research that specifically examines the strategies used to mitigate age-related bias in ML models and the effectiveness of these strategies. Objective To address this gap, we conducted a scoping review of mitigation strategies to reduce age-related bias in ML. Methods We followed a scoping review methodology framework developed by Arksey and O’Malley. The search was developed in conjunction with an information specialist and conducted in 6 electronic databases (IEEE Xplore, Scopus, Web of Science, CINAHL, EMBASE, and the ACM digital library), as well as 2 additional gray literature databases (OpenGrey and Grey Literature Report). Results We identified 8 publications that attempted to mitigate age-related bias in ML approaches. Age-related bias was introduced primarily due to a lack of representation of older adults in the data. Efforts to mitigate bias were categorized into one of three approaches: (1) creating a more balanced data set, (2) augmenting and supplementing their data, and (3) modifying the algorithm directly to achieve a more balanced result. Conclusions Identifying and mitigating related biases in ML models is critical to fostering fairness, equity, inclusion, and social benefits. Our analysis underscores the ongoing need for rigorous research and the development of effective mitigation approaches to address digital ageism, ensuring that ML systems are used in a way that upholds the interests of all individuals. Trial Registration Open Science Framework AMG5P; https://osf.io/amg5p

show abstract

“…Clear annotation manuals with straightforward and transparent instructions are essential to reach valid and unbiased conclusions. A study by Ganel et al [79] showed that the bias related to human assessment is transferable to AI algorithms for automated age prediction, causing poor performance.…”

Section: Label Biasmentioning

confidence: 99%

How artificial intelligence adopts human biases: the case of cosmetic skincare industry

Georgievskaya,

Tlyachev,

Danko

et al. 2023

AI Ethics

View full text Add to dashboard Cite

The cosmetic skincare industry is a growing market that extends to different regions and customer groups. In addition to scientific advances and technological developments, state-of-the-art digital approaches, including machine learning and other artificial intelligence (AI)-based techniques, are being applied at different stages of the value chain. The objectives of these efforts include optimizing the supply chain, developing high-quality, effective and safe products and personalization at every step of the customer journey. However, the use of digital technologies comes with risks and undesirable effects. These include a lack of transparency and accountability, compromised fairness and a general deficiency in data governance, all of which are critical at every customer touchpoint. This dark side of digital transformation is recognized by both businesses and governments. In this paper, we explain the concept of bias leading to unfairness for beauty technology applications. Based on published data we identified potential sources of AI bias in the cosmetic skincare industry and/or beauty tech. They were classified by the stage of the AI lifecycle: biases related to target setting, to acquisition and annotation, to modeling, to validation and evaluation, and to deployment and monitoring. We aim to create awareness of such phenomena among readers, whether executives, managers, developers or potential end-users.

show abstract

Biases in human perception of facial age are present and more exaggerated in current AI technology

Cited by 4 publications

References 31 publications

Genetically Proxied Autoimmune Diseases and the Risk of Facial Aging

Genetically Proxied Autoimmune Diseases and the Risk of Facial Aging

Strategies to Mitigate Age-Related Bias in Machine Learning: Scoping Review

How artificial intelligence adopts human biases: the case of cosmetic skincare industry

Contact Info

Product

Resources

About