South Africa to probe LinkedIn over user data violations

LinkedIn faces probe following accusations of violating user personal information through its new data policy, according to the South African Artificial Intelligence Association (SAAIA).

In an effort to promote responsible use of AI in South Africa, the country founded the SAAIA which is a regulatory body that oversees anything to do with related technology.

LinkedIn unlawfully use personal data to train AI

The social media company is being accused of unlawfully using clients’ data to modify the capabilities of its AI models as per submissions by the SAAIA to the South African Information Regulator.

LinkedIn, in a bid to better its AI abilities is reportedly using the information obtained from its clients to modify their intelligence models to write messages as well as deliver posts on the social media platform on behalf of users.

The AI-developed text will include names of mentors, and work experience gained by LinkedIn on a particular user’s profile on the platform.

The processing of client information is being necessitated by a new LinkedIn policy that enables the platform to process data uploaded by users to improve the AI models automatically. To avoid having their data being used, users have to opt-out.

SAAIA head of regulatory affairs and author of the complaint, Nathan Ross Adams said their submissions are targeting the processing of South African citizens’ data by LinkedIn for the purposes of training their generative AI models.

“The matter requires an investigation from the Information Regulator given the ‘significant public interest’ in the matter.”

Adams

In its submissions to the Information Regulator, the SAAIA said the new policy which enables the social media platform to process users’ data to further develop AI models without permission is prohibited by the Protection of Personal Information Act (PoPIA).

Chapter 3 of PoPIA provisions, in accordance to some conditions, speaks to the matter that personal information processing by a third party is only allowed after consent by the relative owner of the information.

LinkedIn has been caught on the wrong side before

Speaking to News24, Information Regulator spokesperson, Nomzamo Zondi confirmed receipt of the SAAIA complaint and said they are looking at it.

“Once we assess the submissions made by the SAAIA in their complaint, we will communicate our decision to the parties involved, SAAIA and LinkedIn. So, the regulator is not able to make any preconceived notions before conducting a thorough investigation,” said Zondi

Speaking to the same publication, a LinkedIn representative said their platform gives users a choice on how their data submitted on the platform will be used and they are also allowed to not consent to submit their information for AI training.

“We have always been transparent with users on their choices for data submitted to the platform, as LinkedIn has always developed its products through some degree of automation.”

LinkedIn Spokesperson.

The spokesman added that the obtaining reality is that users come to LinkedIn in need of perfecting their draft resume and also to get help from the platform on a cover letter that woos recruiters in order to land their next big career move.

However, the spokesperson failed to talk about the issues of consent as alleged. This is not the first time that the Microsoft-owned LinkedIn is caught on the wrong side of the law. In a similar complaint, the Information Commissioner’s Office, which is the personal information authority in the UK successfully pressured LinkedIn into agreeing to stop the use of UK user data to further develop their AI models.

In other jurisdictions, social media platforms have faced prosecution and regulatory pressure, with Meta and X facing probes in Europe after they were channeling user data to train AI models. This also comes as social media platforms are turning to huge amounts of user data on their platforms to train their AI models.

According to the founder of SAAIA, Dr Nick Bradshaw his association is assessing if the delivery of AI tools to the market is being done in a responsible manner.

“Impacts of AI can be local despite the race for AI development being global. SAAIA has been watching carefully the speed at which AI innovations are being offered by vendors as investors are pumping in huge sums of money to beat opponents to the market, but SAAIA has to ensure the safety of local users,” said Dr Bradshaw.