The world continues to be marveled by the constant rise in new technologies and the advancements in generative AI, which have been a part of human lives over the past few years. And while we continue to make use of these AI tools, which consist of filters and other face apps for fun, we may not have paused to actually look at the untold effects of these tools, some of which are TikTok, Instagram, and Snapchat filters, FaceApp, Lensa AI, etc.
AI beauty filters, which overlay “perfection” on photos, have become prevalent on social media, attracting users with their stunning features. However, these filters also carry risks and implications that should be considered by all.
A very recent AI wave was the Ghibli trend, where you share a photo of you or anything to Chat GPT and it turns it into a cartoon-style image for you. This went viral not just because it was a new trend, but people found it cute. But did we stop for a moment to wonder what happens to our photos after we upload them? Let’s uncover the implications of these AI tools on users.
How these AI filters work
Beauty filters use machine learning algorithms and computer vision technologies to identify and map facial features, after which digital layers are applied to a user’s face to smooth skin, contour face shape, resize features, or apply virtual cosmetics. Advanced filters can also be used to modify lighting and color balance in professional photography.
Once detected, the filters make adjustments based on their functionality, allowing users to create a personalized version of their face. This procedure takes place instantaneously, delivering real-time results for evaluating and tweaking before sharing the final product.
Modern AI learns by example, becoming smarter with more examples. Every time you use filters or generate AI portraits, you contribute to the data pool, helping machines understand race, gender, emotion, and health indicators. This knowledge can be used for good or weaponized in predictive policing, hiring algorithms, and digital advertising.
Dangers of AI filters
Crooked perception of the world and of self
AI algorithms and huge language models are trained on large datasets, yet they frequently contain biases, which lead to biased decision-making. This can gradually change user behavior, beliefs, and personality. For example, a facial recognition algorithm that has been trained on lighter skin tones may struggle to reliably identify deeper skin tones. These algorithms’ predictions can have subtle effects on users’ beliefs and self-perception. The effect continues even after people stop utilizing AI services. To make informed judgments, people must understand how algorithms impact their ideas and behavior patterns. Overuse of beauty filters may lead to body dysmorphic disorder, a mental health disease in which people obsess about perceived defects in their looks that others may not see.
Breach of privacy
The AI-driven ecosystem has resulted in a severe loss of privacy, as users’ time and attention are utilized to collect and analyze personal data. This information is frequently monetized through third-party sales or targeted advertising on social media. Users must be aware of these hidden costs to make informed judgments about the trade-offs between convenience and personalization provided by AI-powered services and solutions. Furthermore, there is a risk of sharing sensitive personal data more broadly than intended, resulting in potential privacy breaches. For example, AI-powered healthcare chatbots collect sensitive health information that could be compromised in a security incident. The AI-driven business model favors data gathering over transparency, undermining the fundamental right to privacy in human interactions.
A tap into human intellect
AI chatbots, LLMs, and content producers frequently keep user data without their explicit agreement, helping suppliers to enhance AI models. However, this data can be used to violate intellectual property rights. Generative AI models enable users to generate original content without their explicit authorization, while the AI creator retains ownership. Engaging with proprietary AI systems entails providing considerable personal data and intellectual property to tech businesses without transparency. This can result in accidentally releasing creative ideas and information that helps AI businesses rather than users.
It is imperative to be mindful of the filters we use. Before using AI photo apps, understand the rights you’re giving up, use local processing for safer apps, push for stronger biometric data protection legislation, and use the option to request deletion.
