Q&A with IBM CEO Arvind Krishna on OpenAI's ChatGPT, research into LLMs, companies using AI, Deep Blue and chess, practical uses for generative AI, and more
Richard Waters / Financial Times : Tweets: @carlquintanilla and @hare_brain Tweets: Carl Quintanilla / @carlquintanilla : IBM CEO: “We do have a shortage of labour in the real world and that's because of a demographic issue that the world is facing. .. So maybe we can find [A.I.] tools that replace some portions of labour, and it's a good thing this time.” $IBM @FT #AI https://www.ft.com/... Stephanie Hare / @hare_brain : Arvind Krishna: If AI can replace labour, it's a good thing IBM's chief executive sees practical use cases for artificial intelligence and quantum computing in just a few years https://www.ft.com/...
Context & Ripple Effects
This interview sits at a hinge in IBM's AI story. It comes after insiders exposed Watson's corporate-market missteps and IBM's retreat to less ambitious commercialization, and months before Krishna rolled out WatsonX — the enterprise platform he would defend later that year in another Q&A on AI's business uses and Biden's executive order.
What makes the piece durable is its economic framing: Krishna argues that demographic-driven labor shortages make replacing portions of labor with AI "a good thing," with practical generative-AI and quantum use cases arriving in just a few years — a thesis IBM then spent the following years testing on itself.
First-order effects
- Krishna publicly stakes IBM's position on AI-as-labor-replacement, tying the company's AI strategy to workforce demographics rather than cost-cutting alone.
- The interview pairs near-term generative-AI use cases with IBM's quantum research, setting expectations that both deliver practical value within a few years.
Second-order effects
- Execution lags the rhetoric: by late 2024, sources report IBM's AI isn't up to meeting Krishna's ~7,800-staff replacement commitment, exposing the gap between the demographic thesis and the product's capability.
- Enterprise buyers watching IBM's shortfall get a calibration point for what current LLMs can actually absorb in back-office roles versus what CEOs claim.
Third-order effects
- By 2025 the thesis lands selectively — IBM deploys AI agents to do the work of 200+ HR employees while hiring more programmers and salespeople ([[a:885429]]) — pointing toward redeployment rather than net headcount reduction as the operating model.
- Krishna was still arguing the same case in 2025, including why there is no AI bubble, suggesting labor-shortage framing has become the durable political-economic justification for enterprise AI adoption.
The trend: Enterprise AI adoption is increasingly justified through labor economics — executives recasting automation as the answer to demographic workforce shortages, with redeployment rather than pure cuts as the proof case.