In November 2023, a video of Indian performer Rashmika Mandanna went viral on social media, where she seemed to be wearing revealing clothing. However, this was a fake video.
It was a deepfake.
A deepfake is AI-generated content that’s a manipulated video or image. The goal is to convincingly replace the likeness of one person with another. So, deepfakes are a form of AI.
You may have certain videos of Hollywood performer Thomas Cruise on social media being light-hearted. There may be a chance the videos you’ve seen aren’t really of him, they’re deepfakes.
In the case of Rashmika Mandanna, the video actually was of British-Indian influencer Zara Patel.
Amidst all the bouquets and brickbats that AI has received, could this be something in the “cons” column: exacerbating harassment of women through AI?
Since deepfake tech could be used to create explicit and fake content by superimposing a woman’s face onto explicit material without her consent, this could be damaging when the manipulated content is shared online or among people. And that could lead to harassment, embarrassment and potential harm to the victim’s personal and professional life. Deepfakes could, also, be used to impersonate women, creating false representations that may involve inappropriate or harmful behaviour. And that could lead to social, psychological and professional consequences.
And it’s not like deepfakes were just born with the advent of generative AI. They are said to have been around since 2017. Sensity AI is an organization monitoring deepfakes online. According to its report called The State Of Deepfakes 2019 Landscape, Threats & Impact, about 96% of deepfakes were non-consensual s*xual deepfakes. Of those, about 99% were made of women.
Another famous example is of female Hollywood performer Scarlett Johansson. She had previously expressed concerns about deepfakes, after her face had been used for multiple s*x scenes using AI software to create videos.
So, what could truly be done about this? Could there be laws to target deepfake creation and distribution? Should social media platforms implement and enforce strict policies against the sharing of deepfake content without consent?
What if the deepfake content is non-s*xual and is meant in good humour? Hollywood performers Thomas Holland and Robert Downey Jr have been, previously, deepfaked as characters from the 1985 movie Back To The Future. That was meant to entertain people and Holland was said to have reacted to this positively.
In November 2023, the Indian government was said to have issued a reminder to social media platforms about the legal provisions surrounding deepfakes. Offenders may face imprisonment for up to 3 years and a fine of up to 100,000 rupees. India’s Minister Of State For Electronics & Information Technology Rajeev Chandrashekhar was said to have underscored the need for vigilance against the misuse of tech. Can there be legal support and recourse provided for victims of deepfake harassment? And could there be development of advanced deepfake detection tools?
Or are all these superficial measures?
These are all potential solutions to a core issue that may not be currently addressed: why is this happening? Why are people using AI in a manner to denigrate women? Is this just old wine in a new bottle: finding innovative ways to make women feel disempowered? Could this be about providing a sense of power and control to the person generating inappropriate content? Could this be a means of seeking attention or validation? Could it be about having deep-seated biases and prejudices against women and using tech to perpetuate harmful stereotypes to contribute to the objectification of women?
Or could it be a desire for escapism and creating deepfakes aligning with certain fantasies? Or is it the power of anonymity that may embolden people to engage in harmful behaviour? The perception of reduced accountability or consequences might encourage such actions.
Or maybe, it’s about the pure fascination that some people may have with the tech.
If it’s the last couple of reasons, then women may be collateral damage to creators who are disregarding the harm that such deepfakes would cause to these victims.
Maybe, if these deepfake creators were to truly understand the ethical implications of creating and sharing such deepfakes, these harmful actions could be curbed. Or maybe, more empathy needs to be encouraged to help people understand the emotional impact of inappropriate deepfakes on the subjects involved. And that just might foster a greater sense of responsibility and restraint.
Could online communities establish and reinforce positive norms that discourage the creation and sharing of inappropriate deepfakes? Could there be peer influence that plays a role in online behaviour and shaping community norms in a positive manner?
Maybe, there could be awareness campaigns exploring the potential real-world consequences of creating and sharing inappropriate deepfakes. These could illustrate the impact on victims’ lives, relationships and mental well-being. Would that deter people from engaging in such behaviour and would they be stuck in their ways?
What would it take to discourage the creation and dissemination of harmful deepfake content?
