Artificial intelligence and cyberbullying: Rethinking legal frameworks for the modern digital age
A case on deepfakes
The swift expansion and evolution of artificial intelligence (AI) have generated numerous legal challenges, affecting both statutory schemes and court proceedings. Although AI now plays a central role in legal research, drafting and client assistance, the statutes that regulate AI have lagged considerably. In the matter of Le Roux and Others v Dey, the case originated in the North Gauteng High Court located in Pretoria. The defendant was Dr Dey, who at that time served as a deputy principal of a school. The plaintiffs, who were students at that institution, posted a picture that placed the defendant together with the school head in a sexually suggestive manner.
The High Court granted damages in the sum of R45,000, citing that the image amounted to defamation. The Supreme Court of Appeal confirmed the judgment of the North Gauteng High Court, while the Constitutional Court found that the image was both defamatory and an injury to dignity, ordering the applicants to pay R25,000 in damages and issuing an apology to the deputy principal.
Gaps in protection
The Le Roux case is widely cited as the legal foundation in South African law for dealing with AI-driven deepfakes, non-consensual pornography, and impersonation. However, the doctrines and statutes applied in that case are insufficient for addressing contemporary AI tools, particularly where cyberbullying and personal impersonation via deepfake technology are involved.
Deepfake technology uses machine-learning models that study a person’s facial features, movements, and voice patterns and can then recreate them in different scenarios. Although the Cybercrimes Act 19 of 2020 and the Protection of Personal Information Act 4 of 2013 may be invoked to sanction certain online conduct, the emergence of AI-driven deepfakes calls for a more robust, full legislative framework.
Victims of cyberbullying, particularly through the use of deepfake AI technology, are often left with no effective legal remedies. Offenders often remain unidentified, employing fictitious accounts and private VPNs that hinder detection. By the moment a court considers an urgent motion, the aggrieved party may already have endured irreversible injury. Identifying a deepfake may require forensic analysis and cooperation with foreign service providers, who are mostly social media platforms. The prolonged period between instituting court processes and engaging service providers allows any harm perpetrated, whether in the form of a fake image, video, or defamatory statement, to spread across cyberspace, causing irreparable damage to victims.
