In this insightful blog post, we explore the emerging phenomenon of deepfakes and their profound implications for the fashion industry. Delving into how advanced AI technology is used to create hyper-realistic digital manipulations, we discuss both the potential benefits and the significant threats posed by deepfakes

DIYA MEHTA

7/8/2024, 4 min read

Lately, the term ‘deepfakes’ has been a catchphrase. But what are deepfakes? Deepfakes are digital media: video, audio, and images edited and manipulated using Artificial Intelligence (AI). The process involves training a machine learning model on a large dataset of images and videos of a target person and then using this model to generate new content that appears to feature the target person. The visual content is virtually modified to drastically change a person’s appearance to make them seem different or look like someone else. The manipulation is to the extent that it produces a convincing fake. Surprisingly, deepfakes in today’s time can be easily generated by semi-skilled and unskilled individuals by morphing audio-visual clips and images. This is because the apps and tools used to manipulate digital media are now easier, faster, cheaper, and more accessible.

The advent of deepfakes has challenged our imaginations as, in today’s time, a lot can be done via dynamic voiceover, face-swapping, etc. Technology has come in handy for the fashion industry. For instance, a Chinese tech company, Tencent, used AI to analyse Chinese teens’ fashion trends. Not only this, but the fashion industry has also been witnessing the use of AI to show outfits in a much wider variety, which includes skin tones, heights, weights, etc., thus promoting consumer engagement. It has also allowed retailers to give customers access to virtual trial rooms. Virtual trial rooms enable customers to use their faces on digital models and analyse whether a particular outfit suits them. This has reduced the need for physical samples and photoshoots, reducing waste and costs. Therefore, AI has changed our shopping experience. From being able to try a shade of foundation virtually to suggesting items we might like based on our searches and previous purchases, it has challenged what we would have comprehended in the past. The use of deepfakes in these applications has significantly enhanced the consumer experience and reduced the fashion industry’s environmental impact.

Job insecurity and being replaced have always been a concern among youth. However, being replaced by a virtual model or an influencer takes things to a different tangent. It is pretty difficult to stomach being replaced by hyper-realistic models and influencers who never age, get tired, or have scheduling conflicts (as they can be present at multiple locations at a given time). Brands have been using virtual models and influencers to promote fashion items. Vogue, for one, included AI models in its 2023 Singapore Issue. Balmain has also introduced a “New Virtual Army” campaign focusing on virtual models. Virtual influencers are also on the rise. They have built a considerable following that could give competition to real influencers. Lila Miguela is one such virtual influencer, built with AI and claims to be 19. She (or it) has almost 2 million followers on Instagram. Besides collaborating with brands like CK and Samsung, she even collaborated with Prada during Milan Fashion Week, “wearing” pieces from the latest collection. If consumers are not made aware of the virtual models, this may lead to distrust in the brands. The rise of virtual influencers and models can hit a blow on the job opportunities for real models. Furthermore, the use of deepfakes in fashion advertising can perpetuate unrealistic beauty standards and create body image issues, as seen in the case of Skinny Tan Ltd’s advertisement.

Fashion shows haven’t been spared either. Fashion shows can feature both real and virtual models. Undoubtedly, this allows designers to showcase their collections innovatively; the possible downside of it must be addressed. For example, the Met Gala 2024 was attended by usual celebrities; some were absent, and some made digital appearances. Katy Perry is one such celebrity who even shared her AI-generated deepfakes to the gram with a screenshot of her mother’s reaction, who believed her AI image to be accurate. Therefore, deepfakes have allowed to blur the line between reality and fiction, leading to a loss of authenticity. The image aligned with this year’s theme, making it more unmissable. Met Gala is an event where fashion is displayed, and designers take inspiration for commercial production. This can hamper creativity in general.

It has also allowed aid personalised marketing by integrating user’s likenesses into fashion promotions. Modelez used Shahrukh Khan for Cadbury by creating his deepfake. The campaign allowed local shopkeepers to create a free personalised ad in which Khan’s deepfake would talk about specific stores based on the information provided by the store owners. Therefore, for a small retailer with a restrictive marketing budget, deepfake offers an easy solution.

Deepfakes can cause body image issues by perpetuating unrealistic beauty standards and creating perfect and unattainable images. For example, Skinny Tan Ltd’s advertisement featured Instagram stories of influencers who had applied filters, making their skin toner darker than it would have been without the filters. The filters were directly relevant to the performance of the products, as the advertisement was for tanning products.

The more severe repercussions of deep fakes can be using technology to impersonate individuals, leading to identity theft and financial fraud. A model’s likeness may be misused, leading to reputational damage and emotional distress. However, the potential for misinformation and manipulation can grow exponentially, especially in fashion, where things are primarily visual. This underscores the gravity of the issue and the need for comprehensive solutions.

India lacks specific laws to address deepfakes. However, the current landscape does offer civil and criminal remedies under different legislations. For example, Section 66E of the IT Act covers capturing, publishing, or transmitting personal images in mass media, thereby violating their privacy and making the offence punishable with up to three years of imprisonment or a fine of Rs 2 lakh. Section 66D of the Act punishes individuals who use computer devices with malicious intent, leading to impersonation or cheating. Furthermore, sections 67, 67A, and 67B can be used to prosecute individuals for transmitting or publishing deepfakes that are obscene. IPC provisions like Section 509, section 499, and Section 153 (a) and (b) can curb the misuse of deepfakes.

Even though specific remedies exist to prevent the misuse of deep fakes, the need for particular legislation cannot be contested. This is because the existing laws are not adequate and cannot be said to be designed to meet the challenges offered by new technologies. Addressing this issue through specific legislation is paramount, and we must act swiftly to protect our society and the fashion industry.