Plan to regulate AI and misinformation in South Africa
In its annual performance plan for 2024/25, South Africa’s Film and Publication Board (FPB) revealed that it is conducting “deep research” into generative AI to prevent the technology from spreading misinformation.
Citing the World Economic Forum, the FPB says misinformation and disinformation are among the most severe global risks anticipated to emerge over the next two years.
The spread of such information can further widen societal and political divides globally.
It also poses a significant threat in South Africa concerning the upcoming national and provincial elections.
“To ensure elections free from misinformation and disinformation and mitigate the consequential risks associated with them, the FPB will support the Independent Electoral Commission by monitoring the distribution of misinformation and disinformation,” it stated.
“This is critical to ensure that the legitimacy of this hard-fought-for democratic process is not undermined in the country.”
It believes that the misuse of generative AI on online platforms could boost the spread of misinformation in South Africa.
Its annual performance plan includes a Strength, Weakness, Opportunity, and Threat (SWOT) analysis, listing its research into generative AI as a strength in its fight against misinformation.
“Deep research on the rise of generative AI to develop measures that can be put in place to prevent the technology’s misuse (misinformation and disinformation),” the strength reads.
However, it also noted the capacity constraints faced by its research units as a weakness.
It plans to turn the weakness into an opportunity by collaborating with stakeholders, including the South African Artificial Intelligence Association, to develop regulations for the use of AI on online platforms.
The FPB has taken several approaches to fight the spread of misinformation in South Africa.
Most recently, it scrapped a policy that would have forced Internet service providers (ISPs) and social media platform owners to police their content for disinformation and misinformation or face criminal penalties.
Under the regulations proposed, ISPs could face fines up to R750,000 and five years in prison for failing to notify the FPB within 30 days of becoming aware that their services were used to spread disinformation or misinformation.
It first published a notice detailing the policy on 22 March 2024, after which it faced legal threats from several advocacy groups, including Media Monitoring Africa and the South African National Editors Forum.
They raised several concerns and threatened to take the FPB to court if it didn’t withdraw the notice by 17:00 on 4 April 2024.
The group were primarily concerned that the FPB’s definition of misinformation was too broad as it implicated people who unwittingly said something false.
They also argued that the regulator had no mandate to regulate misinformation or disinformation, meaning it was trying to enact new legislation without following the correct procedure.
Cliffe Dekker Hofmeyr consulting media and law specialist Emma Kingdon said the FPB was effectively trying to declare that misinformation and disinformation are forms of propaganda for war, incitement of imminent violence, and hate speech.
The FBP withdrew the controversial rules in a notice published in the national government gazette on 26 April.