Once AI bubble pops we'll all suffer – better than letting it grow unabated? (www.theguardian.com)

The piece warns that the global economy is increasingly propped up by a narrow AI investment boom that could be a fragile bubble. Surging capital into a handful of tech giants (Nvidia, Alphabet and other “magnificent seven” firms now worth about a third of the S&P 500) is sustaining stock markets and consumer spending even as wages, hiring and consumer confidence weaken. Market signals such as a rising VIX and investor nervousness suggest a sudden reassessment could erase trillions in household wealth — analysts cite dot‑com–scale losses (Gopinath’s $20tn estimate U.S., $15tn global) — triggering a deep recession if the AI-driven productivity gains investors expect don’t materialize. Technically and socially, the article highlights a crucial distinction: today’s AI push aims not merely to automate specific tasks but to replicate human capabilities at scale, potentially concentrating wealth and decision‑making power in the hands of technology owners. That raises existential labor-market questions about how people will earn livelihoods and whether redistribution can or should compensate. The piece argues for a reorientation: prioritize AI that augments human skills (examples: protein design, giving nurses enhanced diagnostic capabilities) rather than wholesale replacement. It also notes historical precedent—past bubbles left durable infrastructure and knowledge—so a corrective crash might, paradoxically, create space to rebuild AI toward broader, more equitable human benefit.