Listen to the audio version of the article
Revenge porn, sextortion and deepnude, these are the forms of digital violence against women made more accessible with artificial intelligence. The creation of false pornographic content, for example, has found in AI a tool for rapid and dangerous production. What are the limits of application of AI and how to use it against gender violence? Chatbots and apps, handbooks and online services for data protection are only the initial tools for building a long-term best practice.
La cyberviolenza
Some cyberviolence phenomena appear as a crime, others are traced back to it, and this is where AI encounters its first obstacle. Revenge porn has been a crime since 2009 thanks to art. 612 ter of the Italian Penal Code. The non-consensual dissemination of intimate material, online and on Telegram/Whatsapp groups, is growing and also involves minors. Furthermore, explicit content can also be used for extortion purposes, hence the name sextorsion.
The non-profit association Permetto Negato, specialized in supporting victims of non-consensual pornography, reports the increasing cases on Telegram: 147 groups and channels monitored, 16,883,722 users and over 6,704,050 contents shared. The latest analysis at the end of 2023 included a focus on 7 groups, with 331,837 messages involved. The demographics are predominantly male (11-60 years). In the first eleven months of 2023, the Postal Police recorded 259 cases (183 reported by women). How to defend yourself? The Privacy Guarantor provides a handbook on security and offers the possibility of making a preventive report online. In February, on the occasion of Safer Internet Day, Meta launched the Take it down service to block or prevent the spread of private material.
The abuse of AI
What if the explicit content concerns us but is fake? Manipulation for pornographic purposes, or deepnude, makes up 90%-95% of all falsified content, and 90% is porn that victimizes women (Sensity AI). While waiting for updates to the AI Act, reference can be made to further crimes such as defamation or extortion, as well as to the regulations of the social networks involved. Together with videos, AI facilitates the creation of false identities (catfishing), encouraging romance scams: fake profiles that establish an intimate relationship with the scam victim. According to the latest data from Cyber Guru, a security awareness training platform, 52% do not distinguish real profiles from fake ones and 13% have exchanged material with strangers. The danger is coming across a cyber criminal capable of installing malware on your mobile phone and taking possession of your personal data.
The Postal Police recorded a 118% increase in romance scam cases in 2021 compared to 2020 (record year) with sums stolen amounting to 4,500,000; while in the first eleven months of 2023, cases of sextortion rose to 1,384, with a 29% increase from 2022. Catfishing is not a crime but can be traced back to the crime of impersonation or identity theft, regulated by art. . 494 of the penal code. In February, the dating app Bumble released the first data from Deception Detecotor, a system that tracks fake profiles using AI: in the first two months, reports were reduced by 45%, automatically blocking 95%.