- South African businesses are embedding AI into hiring, contracts and customer systems, but many do so without legal safeguards, creating unseen exposure to lawsuits and penalties.
- Automated decisions still carry human accountability, meaning companies remain liable for discrimination, privacy breaches and flawed contracts produced by AI tools.
- Legal advisers say digital transformation must be matched by legal transformation, including audits, stronger governance and updated compliance frameworks.
South African businesses are embracing artificial intelligence and automation at unprecedented speed, weaving algorithms into recruitment, customer service, document drafting and daily operations.
The promise is efficiency and growth. The danger, legal experts warn, is that many companies are moving faster than the law and faster than their own risk controls.
Cor van Deventer, director at VDM Incorporated, says he is seeing a growing pattern of organisations adopting AI tools without first asking what legal responsibility comes with them. In his view, technology is no longer just an operational upgrade. It has become a legal inflexion point that directly affects rights, compliance and accountability.
“AI isn’t just a technology trend, it’s a legal turning point,” Van Deventer says. “Businesses are adopting automated tools without realising that they’re creating new liabilities, new compliance obligations, and new exposure to litigation. The risks are not theoretical anymore. They’re already here.”
He explains that many companies treat AI like software that simply improves productivity, yet the moment it starts making decisions about people or money, it steps squarely into legal territory. Whether it is screening job applicants, approving customers, generating contracts or responding to complaints through chatbots, each automated action can have real-world consequences.
You cannot blame the algorithm
One of the most serious misconceptions, Van Deventer says, is the belief that responsibility can be shifted to technology. If an automated system rejects a candidate, denies credit or flags someone incorrectly, the company remains legally accountable.
“You can’t blame the algorithm, because the law will look to the company, not the code,” he says. “If an automated system makes a decision that affects a person’s rights, the business remains accountable.”
This is especially risky in human resources, where screening tools may unknowingly filter out candidates based on hidden or biased data patterns. Employers may not even understand how the system reached its decision, yet they could still face claims of discrimination or unfair labour practices.
Privacy risks are equally concerning. AI tools often process large volumes of personal information, sometimes pulling in data from third parties or cloud services. Van Deventer says many businesses cannot clearly explain how that information is collected, stored or used, which places them at risk of breaching data protection obligations. If a company cannot justify its data practices, it may already be out of compliance.
Contracts, fraud and ownership questions
The rush to automate has also reached legal drafting. More businesses are using AI to prepare contracts, policies and internal documents. While this may appear efficient, Van Deventer warns that these texts often contain ambiguous clauses, missing protections or incorrect legal references that only surface when something goes wrong.
“AI can assist with drafting, but it cannot replace legal expertise,” he says. “A flawed contract is a liability waiting to happen.”
Cybercrime is evolving alongside these tools. He notes a rise in deepfake voice calls, fake attorney emails and impersonation scams designed to redirect payments or steal funds. Without strict verification procedures, companies may find themselves liable for significant losses.
Intellectual property presents another grey area. AI-generated content raises difficult questions about ownership and originality. Businesses that publish or distribute material without knowing its source could inadvertently infringe someone else’s rights, exposing themselves to disputes and claims.
For Van Deventer, the broader problem is structural. Many organisations upgrade their technology but leave their legal frameworks unchanged. Contracts, policies, compliance systems and governance controls often lag behind the tools being deployed.
“Technology is moving faster than regulation, but that does not protect businesses from liability. If anything, it increases it,” he says. “Digital transformation has to be matched with legal transformation.”
He urges companies to conduct legal readiness audits, strengthen data protection compliance, update contracts, introduce fraud verification protocols and train staff on AI-related risks. In his view, the businesses that thrive will not be those that automate the fastest, but those that build strong legal foundations first.
“AI can transform a business,” he says, “but only if the legal foundations are solid. The companies that succeed will be the ones that innovate responsibly.”
Get your news on the go. Click here to follow the Conviction WhatsApp channel.
