The firms or industries that adopted AI so soon and retrenched their employees are regretting their actions/mistakes. They admitted that they were wrong about the positive possibilities of adopting AI (www.hrdrive.com, 2025).
An AI agent called Claudius was given a job of managing a vending machine for the entire month by Anthropic AI (the founder of Claude chatbot). AI failed to manage the vending machine in a dismal manner. It was tasked with the stocking for the shelves ordering of shop items from suppliers through an email. This shop was based in San Franscisco (USA). Claudius would be cheated by employees of Anthropic. Claudius will let customers buy items with less than the actual price and give freebies to the customers. Claudius also reduced the quoted price given to customers and hallucinated fictious account for customers to pay into it (www.euronews.co, 2025).
Fifty-five per cent of the UK’s firms regret adopting AI and letting go of their workers. This adoption of AI led to the remaining employees resigning and also led to a reduction in production output and quality (https://astuterecruitment.com, 2025).
IBM replaced 8000 workers with AI in 2023 in the Human Resources department. It started using AskHR, an AI agent tool, to reduce labour costs. It was a successful initiative, but 6% of the tasks given to AskHR still wanted a human touch like empathy (http://thedailygalaxy.com, 2025). This IBM scenario means AI is not a panacea for everything and cannot compete with humans in some domains.
A company called Klarna had automated its customer service and marketing tasks by employing AI agents. It retrenched employees using OpenAI software and or its applications. The company just cried aloud, saying it needs its employees back at work. The firm now wants its employees back because AI cannot master all or some of the tasks better than human beings (www.futurism.com, 2025).
Researchers have found that AI has produced false diagnoses when it evaluated medical advice or diagnosis. The research indicates the dire consequences for patients’ quality of medical care (www.ndtv.com, 2025). These medical centres must ensure that AI skills are adopted to harness or supplement human medical skills or tasks, not vice versa.
AI will not replace humans or erode human employment, but it can be integrated into the workplace to enhance production quality. AI will never have human elements like compassion, understanding, sensitivity, collaboration, teamwork, and empathy. AI’s limitations are those of the human touch or emotional intelligence and the understanding of humaneness as an ICT application. AI can assist employees in gathering information, analysing information, suggesting patterns and possible or probable predications of what lies ahead for human beings (www.uxmag.co.za, 2025).
These firms that replaced employees with AI have lost the most common elements of human thought, namely:
- AI does not have the experience to argue with demanding clients.
- The human touch and skills are missing from AI agents.
- AI cannot make quality decisions because clients’ needs and problems vary.
- AI cannot understand the cultural dynamics of clients when interacting with their needs (https://nextgencareers.co.za, 2025).
CONCLUSION: Trust AI at your peril; it is not a panacea for all problems.
By
Dr Sefoko Ramoshaba
