The Guardian
RSS FeedCompanies rushed in on AI. Now some are pulling back
Original Published: October 14, 2026
๐ฏ Impact Sentiment: Concerning
๐ Summary
- Employers globally are pulling back from rapid AI adoption after discovering that rushed deployment creates floods of errors, low-quality output, privacy and security risks, and rising costs that erode productivity gains โ with 90% of executives reporting AI efforts have yet to translate into measurable productivity gains, according to an Atlanta Federal Reserve working paper.
- Meta scaled back plans to replace staff with AI agents after employees spent increasing time cleaning up technical and security errors, and has reversed its middle-manager cuts, expanding managerial ranks again. Microsoft internally found that "adding agents to a broken process still leaves a broken process."
- An Ernst & Young survey found that 36% of senior AI executives at large companies (over $1bn revenue) reported an AI incident that caused data loss, financial damage, brand damage, or operational disruptions in the past year โ a significant risk profile that is prompting companies to implement restrictions.
- Major firms including Microsoft, McKinsey, Palantir, Booz Allen Hamilton and Nvidia are restricting employee AI use due to security and cost concerns, moving from "deploy and experiment" to "sustained, role-specific training" models.
๐ก JR Insights
- ๐ผ Implication: Workers who can navigate the AI adoption cycle โ not just using tools but understanding their failure modes, risk profiles, and governance requirements โ are becoming more valuable as companies mature from enthusiastic early adoption to disciplined deployment.
- ๐จ Risk: Workers in roles where AI was deployed to replace them may face renewed hiring pressure as companies reverse course โ but this window may be temporary; the pullback is tactical, not strategic, and AI adoption will resume once companies develop better governance frameworks.
- โจ Takeaway: Develop expertise in identifying and correcting AI errors in your domain โ the companies pulling back are not abandoning AI, they are looking for people who can use it responsibly. The ability to audit, validate, and quality-check AI outputs is an increasingly rare and valuable skill.