The fear that AI might \u00a0take our jobs\u00a0 is becoming a \u00a0reality\u00a0. An increasing number of companies are opting to replace human workers with \u00a0automated processes powered by AI\u00a0. Yet, the potential promised by these \u00a0AI agents\u00a0 seems almost limitless. However, the reality is more complex, and some companies have had to reconsider their plans<\/a>.<\/p>\n <\/p>\n Many companies have started to develop plans to replace their employees with AI, particularly in the \u00a0customer service\u00a0 sector. However, according to a survey by Gartner<\/u><\/a>, 50% of businesses are abandoning this path due to the challenges posed by transitioning to a service managed solely by AI. A staggering \u00a095%\u00a0 of executives surveyed stated that they would prefer a \u00a0hybrid approach\u00a0, integrating both AI agents and human agents. In the words of Kathy Ross, head of customer service at Gartner: \u201cAI offers significant potential to transform customer service, but it is not a panacea. The human touch remains irreplaceable in many interactions.\u201d<\/p>\n <\/p>\n Perhaps the most notable example of a company regretting its reliance on AI is \u00a0Klarna\u00a0. In February 2024, the CEO boasted that \u00a0AI was doing the work previously handled by 700 employees\u00a0 they had just laid off, and claimed it was receiving the same customer ratings. A year later, however, the quality of service delivered by AI received considerable backlash, forcing the company to \u00a0rehire human workers\u00a0 to ensure that customers could always speak with a human representative.<\/p>\n <\/p>\n While \u00a0current AI agents\u00a0 may not be as reliable as initially promised by leaders like Sam Altman, it does not guarantee that future iterations won\u2019t improve. Moreover, there are cases where replacing humans with AI has proven beneficial. For example, the CEO of an Indian startup that laid off \u00a090% of its staff\u00a0 reported a year later that this decision was a success<\/u><\/a>. Companies like \u00a0Duolingo\u00a0 have also followed suit, laying off \u00a010% of its translators\u00a0 to replace them with AI. Other major firms like UPS<\/u><\/a> and Cisco<\/u><\/a> have taken similar routes. IBM is a unique case as it has invested in AI while also hiring more workers specifically to manage that AI.<\/p>\n With \u00a0generative AI\u00a0 decelerating, another concept has gained relevance: that of \u00a0AI agents\u00a0. Unlike chatbots, which can handle only one request at a time, an AI agent can manage more complex tasks. For instance, it could organize a vacation simply upon mentioning the dates and destination. The promises of autonomy and multitasking capabilities make AI agents a potential \u00a0threat\u00a0 to numerous jobs.<\/p>\n <\/p>\n Everyone can make mistakes, and AI agents are no exception. If an AI agent makes an error, it tends to repeat it in subsequent tasks, creating a \u00a0compounding effect\u00a0 of issues and resulting in increasingly flawed outcomes. Other attempts to utilize AI tools in the workplace, such as \u00a0disastrous AI-job interviews\u00a0, have not fared much better.<\/p>\n <\/p>\n A study conducted by researchers from Carnegie Mellon University<\/a> revealed that the technology is still in its infancy. They created a fictitious software company with AI agents as employees. Assigning roles such as finance, administration, and software engineering, the outcome was dire: the AI managed to complete only \u00a024%\u00a0 of the tasks assigned to them.<\/p>\n <\/p>\n The impact of AI on the job market is undeniable. According to the latest World Economic Forum report<\/u><\/a>, it is estimated that \u00a092 million jobs\u00a0 will be lost by 2030 due to AI-driven automation. However, there is also a flip side. Concurrently, \u00a0170 million new jobs\u00a0 are expected to be created, with AI being a major driver of job creation, as exemplified by the case of \u00a0Salesforce\u00a0.<\/p>\n As we navigate this complex landscape, it is clear that while AI has the potential to transform industries, it simultaneously raises important questions about the future of work and the necessary balance between technology and human input.<\/p>\nAI Yes, But Under Human Supervision<\/h2>\n
The Klarna Case<\/h2>\n
Not So Fast<\/h2>\n
The Promise of AI Agents<\/h2>\n
Reliability Concerns<\/h2>\n
A Revelatory Experiment<\/h2>\n
AI and the Job Market<\/h2>\n