ARTIFICIAL intelligence may be advancing at a pace that is reshaping entire industries, but it is not displacing the need for human judgment, accountability or domain expertise. Instead, it is reorganizing the structure of work itself, according to Arun Biswas, Global AI and Sustainability Leader at IBM Consulting.
Speaking to The Manila Times after his presentation at the Asian Development Bank, Biswas said the future of employment will be defined less by machines replacing people and more by the emergence of new human roles created by increasingly capable AI systems. “Human judgment and accountability will remain,” he said. “AI is a machine. You can’t hold AI accountable. You have to hold a human accountable at the end of the day.”
Across Asia, governments and industries are wrestling with the implications of rapid AI adoption, from job displacement to workforce readiness to the need for new governance frameworks. Biswas argues that the real challenge is not simply teaching workers new technical skills but preparing them for roles that do not yet exist.
“In any of these rapid technological shifts, you not only need people to learn new skills, you’ll actually have new roles coming up,” he said.
He pointed to software development as a clear example. While AI systems are becoming increasingly capable of generating code, reducing the volume of traditional programming work, this does not eliminate the need for people. It shifts the work toward validating AI‑generated outputs, setting operational context, monitoring performance and ensuring compliance.
“We may not need a lot of people who do coding,” Biswas said. “But we will need a lot of people who can verify and validate that code, who can set the context, who can monitor that code.”
The same pattern is emerging in sectors far beyond technology. In banking, customer service functions are being augmented or replaced by AI‑powered agents. In energy utilities, grid operators are beginning to receive AI‑generated recommendations rather than relying solely on experience or historical practice. Yet in all these cases, the human remains the final decision-maker.
“Tomorrow they will have AI suggest decisions to them and they’ll have to verify whether the decision is correct,” Biswas said. This shift, he added, requires a more deliberate approach to workforce development. Rather than broad, generic upskilling programs, organizations need to map existing roles, identify which functions are likely to evolve and build targeted learning pathways for employees transitioning into future positions.
The urgency of this transition is increasing as AI systems evolve from tools that merely retrieve information to “agentic AI” capable of carrying out tasks with a degree of autonomy. These systems behave differently from traditional software. They are probabilistic rather than deterministic, meaning they may produce different outputs from similar inputs. That variability introduces new governance challenges and raises the stakes for oversight. “It could be right, it could be wrong,” Biswas said. “The treatment of this technology versus other technologies has to be a bit different.” As AI becomes more deeply embedded in enterprise operations, organizations must strengthen auditability, risk controls and decision‑governance frameworks, particularly in sectors where errors carry significant economic or social consequences such as energy, health care, finance and transportation.
Biswas acknowledged that public sentiment toward AI has become more cautious as the technology expands into everyday life. Concerns about employment, misinformation and social trust are rising, though he stops short of calling it a backlash. The challenge, he said, is ensuring that workers understand not only the power of AI systems but also their limitations. “You need people who can make sense of an AI output,” he said. “You need people who can apply judgment.” He compared the emerging human‑AI relationship to the role of radiologists in modern medicine. AI systems can identify patterns in scans with remarkable accuracy, but medical professionals remain responsible for interpreting recommendations and making final decisions. The same principle, he believes, will apply across industries.
As AI becomes more powerful and more autonomous, the workforce will look very different from today’s. But Biswas is clear that the objective is not to remove humans from the process. Instead, it is to build a partnership in which machines provide intelligence and recommendations while people retain responsibility for judgment, oversight and accountability.