Anthropic, a leading U.S. artificial intelligence lab, anticipates that AI-driven virtual employees could begin entering the workplace as early as next year. According to Jason Kuster, Anthropic’s Chief Information Security Officer, these virtual employees will possess individual “memories,” defined job responsibilities within a company, and even independent company accounts and passwords. This level of integration distinguishes them from existing AI agents, which typically function as specialized tools for tasks such as identifying phishing attacks or potential security threats. Virtual employees represent a new category of AI: intelligent systems with broader responsibilities that deeply integrate into daily workflows and can make autonomous decisions. However, Kuster cautions that the adoption of these AI agents will introduce significant cybersecurity risks, including securing AI employee accounts and mitigating the risk of unpredictable behavior. He emphasizes that substantial security challenges remain.

The increasing prevalence of AI colleagues is fueled by companies’ desire for cost savings and efficiency gains compared to human staff. Shopify’s CEO Tobias Lütke recently instructed employees to prioritize AI tools over requesting additional personnel. Klarna’s CEO Sebastian Siemiatkowski has taken an even more assertive stance, asserting that generative AI has eliminated the need for new hires. Siemiatkowski revealed that his company ceased all hiring a year ago, and its AI assistants now manage workloads equivalent to approximately 700 customer service representatives, reducing resolution times from eleven minutes to two.

Beyond cost and efficiency advantages, AI employees may also facilitate innovative working models. They excel at programming tasks, with tech companies like Google and Meta increasingly delegating this workload to internal AI systems. Virtual employees can operate continuously and process vast datasets rapidly, potentially accelerating workflows considerably. However, the rise of AI has also raised concerns about potential large-scale unemployment, with research indicating that administrative and repetitive jobs are particularly susceptible to advanced AI systems.