Of all the technology chief executive officers promoting the potential of artificial intelligence to empower employees, IBM’s CEO Arvind Krishna has emerged as one of the most vocal advocates of its capacity to substitute them.
In the previous spring, Krishna expressed his view that up to 30 percent of IBM’s administrative roles could be automated through AI. Consequently, IBM is among the pioneering companies to join a coalition of prominent technology firms, such as Cisco, Google, Microsoft, Intel, SAP, and others, to address the impact of AI on the workforce. The consortium will also involve advisors from organizations like the American Federation of Labor, DigitalEurope, and Khan Academy.
The primary objective of the group, seemingly inspired by the joint Trade and Technology Council of the United States and the European Union, is to investigate the influence of AI on information and communication technology (ICT) positions. Initially, the consortium will focus on 56 roles that are susceptible to automation by AI. IBM has identified these roles as encompassing 80 percent of the top 45 ICT job designations.
According to the consortium’s plans, they will propose and facilitate training initiatives aimed at assisting students, individuals transitioning careers, and current IT professionals in preparing for and transitioning to roles that are less likely to be replaced by AI models.
Moreover, it seems that the Biden administration is supportive of allowing the very technology companies developing AI replacements to lead this initiative.
“We acknowledge the interconnectedness of economic security and national security,” stated US Secretary of Commerce Gina Raimondo. “I appreciate the participation of consortium members in addressing the emerging workforce requirements resulting from the rapid advancement of AI.”
AI tools designed to enhance productivity, such as Microsoft’s Copilot for Office 365 and Github, Google’s Gemini for Workspaces, and SAP’s coding assistant, have gained significant attention, albeit not always positive, in the past year.
Simultaneously, companies like Nvidia and IBM have been promoting tools to assist enterprises in constructing, refining, and tailoring large language models (LLMs) for internal operations and procedures. IBM introduced its Watson-X platform last spring, while Nvidia launched NIMs, containerized models intended to simplify the development of AI applications and consolidate efforts.
All these initiatives are founded on the belief that AI will enhance the efficiency of workers, enabling them to accomplish more tasks in less time and with fewer resources.
While this concept may be enticing for industries already grappling with staff shortages, the primary concern lies in AI’s potential to reduce the workforce, hence the emphasis on retraining in the announcement.
“The consortium members universally acknowledge the urgency and significance of their collective endeavors with the proliferation of AI across all business domains and the necessity to establish an inclusive workforce offering sustainable opportunities,” as outlined in IBM’s announcement.
Collectively, the group aims to retrain and transition over 95 million IT professionals in the next decade.
Many of these individuals are likely to find themselves in roles related to “prompt engineering” — or as humorously depicted by comedian John Stewart as the “types question guy” — where their responsibility involves formulating instructions used to guide AI systems. However, recent research indicates that AI outperforms humans in generating prompts for AI systems.
At first glance, the idea of retraining workers for an AI-driven world appears responsible. Nonetheless, it is worth noting that the data utilized to evaluate AI’s impact on the workforce could just as easily be leveraged to identify which positions to eliminate first and the speed at which these roles can be phased out without evoking negative perceptions.
Considering IBM’s history of controversial human resource practices, caution is warranted.
This situation mirrors the analogy of the [disproved] boiling frog dilemma. Placing a frog in hot water would prompt it to jump out, but gradually increasing the temperature would lead to its demise. Similarly, moving too swiftly in implementing AI changes could potentially trigger backlash or even destabilize the economy. Conversely, by introducing AI changes gradually, proponents argue that workers will have the opportunity to adapt and transition effectively.
