top of page

The Right to Speak to a Human Is Now Law. Just Not Here

5 hours ago
3 min read

Californians and EU financial consumers can insist on a person, but South Africans have no equivalent protection.



Sign up for my Substack daily AI newsletter here.




In California, a large company that puts a chatbot between you and its service desk must now make a good-faith effort to connect you to a person within fifteen minutes of asking. It must also tell you when you are talking to a machine. Governor Gavin Newsom signed the law in September. In Ireland, under new European rules, consumers buying financial products online or by phone can now ask to speak to a person instead of an AI chatbot. The right to speak to a human is now law. Just not here. As South African banks, insurers, municipalities and government departments automate their front desks, it is time we asked why.


CONTEXT AND BACKGROUND

These laws reflect a simple concern. Automated systems handle typical cases well, but people whose circumstances do not fit the system need someone who can use judgement. Demand for that person remains strong, especially among older customers. In a survey of five thousand American consumers, more than seven in ten people aged fifty-five to seventy-four said they prefer a human agent, a share that is rising.


South Africa’s closest equivalent is narrow. Section 71 of POPIA protects people from decisions with legal or substantial effects based solely on automated profiling. Where such decisions are permitted, people must be able to make representations and understand the logic involved. It gives no general right to reach a person when a chatbot cannot help.


INSIGHT AND ANALYSIS

Michele van Eck, associate professor of law at Wits University, points out that South African law does not even require companies to disclose that a customer is speaking to an AI voice agent. The Consumer Protection Act prohibits misleading conduct but may not reach financial services. The FSCA’s Treating Customers Fairly framework bars unreasonable barriers to help without spelling out a human route. She calls for disclosure at the start of every interaction and the ability to escalate to a human at any point.


The cost of getting this wrong is already visible in the public sector. Research on the Social Relief of Distress grant found that about ninety per cent of exclusions were erroneous, driven largely by automated bank checks. Appeals exceeded ten million in a year, with a ninety-eight per cent rejection rate, and applicants received generic codes with no way to submit new evidence. That is what a system without a human looks like.


I have previously written about this in the context of public services. I argued that when AI becomes the main entry point, bottlenecks move into a digital layer the most marginalised cannot reach, so human escalation must be built in from the start. The same principle applies to private companies. Whether the front desk belongs to a grant agency or a bank, an automated system that cannot hand a difficult case to a person has the same effect: the people with the least power are the ones left without an answer.


IMPLICATIONS

For policymakers, the gap is clear, and the examples abroad show that such a right needs teeth. Public prosecutors can enforce California’s law, with penalties of up to five thousand dollars for a first violation and up to ten thousand dollars for each one after that. A right to a human could sit in consumer protection law and in the conduct rules for financial services. It could start with large organisations and the decisions that matter most: credit, insurance claims, account closures, grants and identity documents. Disclosure that a customer is dealing with AI should be the minimum, and a voluntary code would not be enough.


Business leaders have no need to wait. Guaranteeing a human route, and saying so clearly, is a mark of trust that customers notice. It also protects the company when automated systems fail in the cases that matter most.


CLOSING TAKEAWAY

Automation will keep improving, and much of it will serve customers well. The question is not whether machines answer first, but whether a person is available when the machine cannot help. In a country with deep inequality, many languages and millions of people with limited digital access, that guarantee matters more here than in California or Dublin. A right to a human is not resistance to technology. It is the assurance that judgement and accountability remain within reach. South Africa should write it into law.


Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net



 
 
 

Comments


Leveraging AI in Human Resources ​for Organisational Success
CTU Training Solutions webinar

bottom of page