Your company is already legally liable for what your AI decides. Here is what the law says.
Katlego Sekhu
Artificial intelligence is no longer a future consideration for South African businesses. It is already shortlisting candidates, informing decisions, and in some cases, making calls that would once have required a human being.
The question most employers have not asked is a simple one: when the algorithm gets it wrong, who is responsible?
Kaya BIZ with Gugulethu Mfuphi spoke to Jeffrey Alsop, Associate for Transactional Services, Employment and Incentives at Pinsent Masons, to find out.
According to Alsop, the answer is already clear under existing South African law. The company is.
“The AI, for all intents and purposes, acts as the employer for purposes of certain decisions,” Alsop explains.
That means every candidate it filters out, every outcome it influences, carries the same legal weight as if a person had made that call directly.
If an AI recruitment system filters candidates on discriminatory criteria, the employer faces a potential claim under the Labour Relations Act, regardless of whether a human being ever reviewed the outcome.
When the board delegates to an algorithm
The liability question becomes more complex at the board level. If a director votes in favour of a decision based primarily on an AI recommendation, Alsop raises a question that has not yet been tested in South African courts: did that director actually exercise their power, or did they effectively delegate it to a machine?
It is a fine but important distinction, and one that points to a broader principle: AI can assist human decision-making, but it cannot replace it.
What employers should be doing now
South Africa does not yet have comprehensive AI legislation. But Alsop is clear that the absence of a dedicated law does not create a gap in liability.
He points to the EU AI Act as a useful framework, one that takes a risk-based approach and distinguishes between low-risk uses of AI and high-risk ones, such as systems that determine who gets hired or fired.
His advice to employers is. “Always make AI, to the extent you can, subject to oversight by a responsible person, and that responsible person then takes the final decision.”
To hear the full discussion, listen to the Kaya BIZ podcast.
Read Next: Home Affairs refutes claims of passport-free travel for Lesotho Nationals
The post AI can assist decisions, but responsibility remains with the employer, experts warn appeared first on KAYA 959.
