Artificial Intelligence Beauty Filters and Aesthetic Surgery: Insights from TikTok’s Bold Glamour Filter

Abstract

Background:

Artificial intelligence (AI)–driven facial modification technologies, such as TikTok’s Bold Glamour filter, have transformed perceptions of beauty and self-image. This study aimed to quantify the facial modifications produced by the Bold Glamour filter using automated and manual measurement techniques and to examine their implications for aesthetic surgery.

Methods:

We used images of 10 anonymized female participants older than 18 years; both the original images and those generated after applying the filter were analyzed. The Face++ Dense Facial Landmarks application programming interface was used for detailed facial feature mapping, alongside manual measurements and observations of facial metrics and soft-tissue changes. Descriptive statistics were applied to evaluate proportional and symmetry differences.

Results:

The filter produced marked facial modifications, including increased nasal tip rotation, thinning of the nasal tip, fuller lips, lift of the right eyebrow, more pronounced malar projection, and increased facial width and jaw prominence. Structural dimensional changes were achieved through small but dramatic alterations, predominantly less than 10%, while retaining a natural appearance and enhancing symmetry. Soft tissue changes were also notable.

Conclusions:

TikTok’s Bold Glamour filter enhances perceived attractiveness through AI-driven facial modifications. Although potentially useful in aesthetic procedure planning, there are concerns regarding unrealistic beauty standards and psychological well-being. Plastic surgeons must balance the benefits of AI-assisted planning with patient well-being and ethical care.

Takeaways

Question: What specific facial changes does TikTok’s Bold Glamour filter produce, and how might these artificial intelligence–driven modifications influence expectations and planning in aesthetic surgery?

Findings: Using Face++ and manual analysis of 10 adult female faces, the filter induced subtle (<10%) dimensional changes, including nasal tip rotation, malar projection, and increased lip height. The most prominent effects were nonstructural soft tissue changes, such as skin smoothing and eyebrow thickening, mimicking nonsurgical rejuvenation or makeup.

Meaning: Although potentially useful for aesthetic planning, algorithmic modifications risk promoting unrealistic expectations, underscoring the need for ethical, patient-centered application of AI in surgery.

INTRODUCTION

In the modern digital age, the convergence of artificial intelligence (AI) and social media has modified perceptions of beauty and self-image.¹ AI-driven facial modification technologies, such as TikTok’s Bold Glamour filter (ByteDance Ltd., Beijing, China), allow users to instantly transform their appearance, reshaping physical self-perception.^2,3 The Bold Glamour filter, used by more than 400 million users since its February 2023 release, exemplifies the evolving landscape of digitally influenced beauty standards.⁴

The psychological implications of social media use on body image are a growing concern among researchers and healthcare providers.^5,6 Detrimental impact of platforms such as Facebook (Meta Platforms, Inc., Menlo Park, CA) and Instagram (Meta Platforms, Inc.) on women’s self-perception and self-esteem provides context for understanding the potential ramifications of the Bold Glamour filter.^6–8

This study quantitatively evaluates the facial changes created by TikTok’s Bold Glamour filter to explore the clinical utility and implications of AI-driven beauty tools in aesthetic surgery. By analyzing facial measurements before and after filter application, we aim to deconstruct the specific alterations produced and provide insights into how such digital enhancements may influence perceptions of attractiveness, patient expectations, and surgical planning.

METHODS

Individuals who expressed interest in participating in our study were included. We used both automated and manual methods to analyze the effects of TikTok’s Bold Glamour filter. Using an iPhone 12 Pro Max’s front-facing camera (Apple Inc., Cupertino, CA), we captured photographs of 10 anonymized female participants older than 18 years, with a range of skin tones, and then applied the Bold Glamour filter to the raw images. These images were analyzed pre- and postapplication of the filter (Fig. 1), using the Face++ Dense Facial Landmarks application programming interface (Megvii, Beijing, China), an online software capable of identifying nearly 1000 facial landmarks on uploaded photographs, with the goal of providing precise, standardized coordinates for various facial features.⁹ This advanced application programming interface allowed for detailed examination of changes in facial proportions, including the hairline, forehead, eyes, nose, mouth, and chin, on the dimensional level of digital pixels (1 pixel = ~0.26 mm). Manual observations and measurements were conducted to validate the Face++ results and to identify soft tissue changes, such as rhytid removal and makeup application. Our analytic focus encompassed general proportional and symmetry percentile adjustments across the facial x and y axes, and a focused aesthetic feature analysis examining specific aspects such as radix to nasal tip, upper and lower lip height, intercanthal distance, eyebrow peak height, and canthal tilt. To interpret our findings, descriptive statistical analysis, including mean, median, range, and SD calculations, was used with data outliers identified and removed before improving accuracy.

Fig. 1.

Participant photograph enhancement using the Bold Glamour filter. Unfiltered images (left columns) show photographs taken with the iPhone 12 front-facing camera before the Bold Glamour filter is applied. Filtered images (right columns) show changes in patient appearance after the Bold Glamour filter is applied.

RESULTS

The effects of the Bold Glamour filter on various facial dimensions were quantified for 10 female subjects using automated and manual measurements. General proportional changes are summarized in Figure 2. Across all facial features, notable increases occurred in both the x and y dimensions after filter application, with mean x changes ranging from 7.25% (mouth) to 8.78% (mandible) and mean y changes ranging from 8.65% (mouth) to 11.13% (forehead). The greatest proportional increases were seen in the mandible in the x axis (8.78% ± 4.13%) and forehead in the y axis (11.13% ± 9.24%).

Fig. 2.

General proportional changes of facial features in x and y dimensions (n = 10).

Specific aesthetic facial feature alterations are shown in Figure 3. Facial width increased modestly with filter use by 1.74% ± 2.57%. Facial height values were augmented by 50.05% ± 10.53%. Multiple nasal dimensional decreases were observed, including reductions in nasal tip width (−3.64% ± 3.96%) and alar base width (−1.94% ± 4.26%). In contrast, substantial gains were found in both upper lip height (+25.30 ± 30.30%) and lower lip height (+16.04% ± 13.48%). Only negligible length differences were observed in the philtral columns (0.17% ± 2.54% increase). The intercanthal distance widened slightly, with values increasing by 1.53% ± 2.72%. Canthal tilt was reduced by a small degree (−0.59% ± 1.19% decrease). Regarding the eyebrows, the right peak-to-hairline distance was lessened by −3.81% ± 7.69%, effectively lifting the right lateral brow. Smaller movements occurred on the left (−1.84% ± 7.84%).

Fig. 3.

Aesthetic feature proportional change after the Bold Glamour filter was applied.

Manual before-and-after observations revealed increased nasal tip rotation; thickening of the eyebrows; reduction of tear trough deformities, nasolabial sulci and rhytids; and smoother skin with the elimination of hyperpigmentation, as demonstrated in Figure 4.

Fig. 4.

Manual pre- and postfilter analysis revealed several consistent aesthetic modifications: increased nasal tip rotation, thickening and darkening of the eyebrows, reduction of tear trough deformities, softening of nasolabial sulci and perioral rhytids, and overall skin smoothing with elimination of hyperpigmentation. A, Unfiltered image taken with the iPhone 12 front-facing camera before filter application. B, Image after application of TikTok’s Bold Glamour filter, demonstrating visible changes in nasal tip rotation, eyebrow definition, and skin texture.

DISCUSSION

This study found that use of TikTok’s Bold Glamour filter results in marked facial modifications, including increased nasal tip rotation, thinning of the nasal tip, fuller lips and eyebrows, lift of the right eyebrow, more pronounced malar projection, and increased facial width and jaw prominence. This study’s findings suggest that both feminizing and masculinizing alterations (eg, eyebrow lifting and increased jaw prominence, respectively) are incorporated into the filter’s algorithm, likely to accommodate both male and female users. These alterations theoretically enhance perceived attractiveness, thus contributing to the filter’s popularity and use among influencers and celebrities.¹⁰

In terms of structural dimensional changes, the algorithms seem to augment facial attractiveness gradually through small but impactful alterations, mostly increasing or decreasing the original facial feature dimensions by less than 10%, retaining natural appearance while enhancing symmetry. The nuanced application of multidimensional analysis powered by AI, followed by subtle, distinct, proportional changes to the facial structure and soft tissues, demonstrates a sophisticated capacity for facial enhancement. However, the most notable effects of the filter seem to be nonstructural soft tissue changes, such as smoothened skin and thickened eyebrows. Notably, skin smoothing seemed to reduce visible rhytids, tear trough deformities, and nasolabial sulci. These features are recognized indicators of facial aging and are commonly targeted in periorbital and midface aesthetic treatments, many of which are achievable through nonsurgical techniques such as facial rejuvenation procedures, makeup application, and microblading. This suggests that the filter may be algorithmically aligned with established rejuvenation goals.

Our findings regarding alterations in proportions with the Bold Glamour filter generally align with what is considered traditionally beautiful for the female face.¹¹ We found that both upper and lower lip height proportionally increased with use of the filter, aligning with prior studies indicating that a 53.5% increase in lip surface area and a 2:1 ratio of lower to upper lip height are perceived as most attractive.¹² Future work should quantify the degree of augmentation observed with filter application and formally assess these proportional relationships. Another study showed that longer faces were deemed more attractive, which was also reflected in our findings.¹³ Other elements such as smile angulation, harmonization of features, and assessment of eyebrow width have been shown to play important roles in the idealization of female facial beauty and merit inclusion in future analyses using AI-driven technologies such as the Bold Glamour filter.¹¹ It would also be informative to examine whether the filter’s alterations align with classical aesthetic principles such as the golden ratio (1:1.618), as this may offer insight into whether AI-generated modifications reflect evolving or established beauty norms.

These aesthetic insights reflect a larger technological shift, as AI is increasingly influencing clinical decision-making in both aesthetic and reconstructive plastic surgery. Its applications span the full surgical continuum, from preoperative planning to postoperative monitoring. These include machine learning–based risk stratification, wound healing prediction, and computer vision-assisted anatomical analysis in real time, underscoring the growing role of prognostic technologies in clinical decision-making across the discipline.^14–16 In the aesthetic domain, recent consensus guidelines have encouraged the incorporation of structured assessment frameworks, such as the Facial Aesthetic Index, Facial Youth Index, and Skin Quality Index, into AI-assisted workflows.¹⁷ Although these indices were not originally developed for AI use, their adoption could improve standardization in photography analysis, treatment planning, and outcome tracking. This movement toward integrating objective, quantifiable metrics into digital tools lays the groundwork for more dynamic, patient-facing applications in aesthetic care.

From a clinical perspective, the use of AI-driven filters such as Bold Glamour may serve as adjunctive tools in preoperative aesthetic consultations. By offering dynamic visualizations of potential postprocedure outcomes, these technologies can enhance patient engagement, facilitate communication, and support individualized planning strategies. The findings of this study corroborate current literature that demonstrates how AI-driven filters can assist in visually modeling outcomes for a variety of aesthetic procedures, including lip and brow lifts, blepharoplasty, tip-defining rhinoplasty, face lifts, and skin resurfacing via chemical peels and laser treatments.^18,19 By providing patients with personalized, algorithm-generated previews based on normative aesthetic data, these tools may help set more realistic expectations and improve shared decision-making during consultations.¹⁸

However, the adoption of such technology is not without risks. Algorithmic modifications might oversimplify individual anatomical complexities and omit critical considerations of surgical, anesthetic, and patient-specific risks.^20,21 These filters could also propagate unrealistic expectations, potentially exacerbating psychiatric conditions such as body dysmorphic disorder, depression, and eating disorders due to altered self-perception.²² Studies suggest a troubling correlation between the use of facial cosmetic procedures and the incidence of psychiatric diagnoses, underscoring the need for thoughtful ethical oversight and clinician-guided counseling about the limitations of algorithm-informed appearances.²³

There are several limitations to this study and our findings. The focus on a singular digital tool and its impact on a smaller sample size of female-only patients highlights the need for broader research encompassing various digital beauty enhancements and their effects across different populations, including men and nonbinary individuals. Manual photography with an iPhone without standardized lens settings or lighting, which caused minor but notable differences in quality and dimensions of the patients’ original photographs, likely impacted the accuracy of the Face++ measurements where certain facial dimensions were unable to be plotted and measured appropriately. To overcome this limitation, we evaluated proportional changes rather than exact measurements and identified data outliers with subsequent removal to standardize our data and compensate for this limitation.

Additional research should incorporate larger sample sizes of more diverse participants with standardized, professional photography. Future studies should also explore the psychological and emotional effects of interacting with AI-modified imagery, particularly among adolescents and other vulnerable populations. Investigations that integrate perspectives from ethics, psychiatry, behavioral science, and computational modeling will be critical to fully understand the clinical implications and guide the responsible use of algorithmic facial analysis and augmentation in aesthetic practice.

CONCLUSIONS

This study quantifies the facial aesthetic modifications created by TikTok’s Bold Glamour filter, highlighting its potential to enhance perceived attractiveness through AI-driven changes. Although such algorithmic tools may serve as adjuncts in aesthetic surgical and nonsurgical procedure planning, these tools raise ethical concerns around promoting unrealistic beauty standards and exacerbating psychological comorbidities. Adopting these technologies in clinical settings requires careful consideration, further research, and guidelines for responsible use. Plastic surgeons must balance the benefits of AI-assisted planning with prioritizing patient well-being and providing ethical care.

DISCLOSURE

The authors have no financial interest to declare in relation to the content of this article.

PATIENT CONSENT

Patients provided written consent for the use of their images.