Firms are scaling back AI use as problems outpace gains
Companies that adopted AI rapidly are finding that the tools can create extra work, errors and security risks, prompting some to restrict their use, retrain staff and bring back experienced workers. The pullback highlights a gap between heavy investment and proven productivity gains, while employees and managers face pressure to use tools without enough guidance.
Meta has eased plans for staff cuts linked to AI agents after employees spent more time fixing technical and security problems, and it is rebuilding some management ranks. Microsoft leaders have warned that AI-generated material can produce repetitive work, while Walmart staff have reported errors and false alerts from workplace systems. An Atlanta Federal Reserve working paper found 90% of executives said AI efforts had not yet improved productivity; an Ernst & Young survey found 36% of senior AI executives at large companies reported an AI-related incident.
- Fast AI rollouts are creating errors and extra work.
- Some firms are restricting tools and rebuilding teams.
- Most executives report no productivity gains yet.
New here? Start with this
Artificial intelligence tools can generate text, analyse information and automate some workplace tasks. Many companies have invested in them to save time and reduce costs, but using them can also require staff to check results, correct mistakes and manage security risks.
Company leaders decide how these tools are used, while employees work with them and often handle problems they produce. Some firms are changing their plans, limiting certain uses or bringing experienced staff back into roles as they assess the results.
The issue matters because companies are spending heavily on AI and making decisions about jobs and working practices before its benefits are clear. Reliable guidance and training can affect whether the tools reduce workloads or add to them.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Companies may be right to scale back or limit AI where it creates more work than it saves, produces errors or exposes sensitive data. The reported experiences suggest that rapid adoption without adequate testing, guidance and oversight can undermine productivity and put staff, customers and the business at risk.
The case against
The problems may reflect early implementation and insufficient training rather than limits that justify a broad retreat from AI. Companies could preserve potential long-term gains by improving oversight, retraining staff and applying the tools selectively, while treating current errors and security incidents as reasons to refine deployment rather than abandon it.
AI Art Business Companies Culture Technology
Read the full article at the source →
Originally published by The Guardian as “Companies rushed in on AI. Now some are pulling back”.