Firms scale back AI spending as executives cite low quality output and costly errors

Major technology firms are reducing their reliance on artificial intelligence tools after discovering that employees are using the systems to produce low-quality work rather than to enhance productivity. Reports indicate that companies such as Meta are scaling back plans to replace staff with AI agents, while Microsoft is also adjusting its approach following internal concerns about the quality of generated content.

Ryan Roslansky, executive vice president at Microsoft, expressed frustration on LinkedIn regarding the prevalence of what he termed low-quality AI output in daily work. He described receiving a document that was clearly AI-generated and noted that summarising it with another AI tool yielded no new insights. Roslansky characterised this cycle as a negative loop that fails to add value to business processes. This sentiment reflects a broader corporate realisation that significant budget allocations for AI tokens are often wasted on efforts that appear productive but lack substantive merit.

Johnny Taylor Jr, CEO of the Society for Human Resource Management, told the Guardian that executives feel they are simultaneously operating and building the AI infrastructure, a situation he described as problematic. He suggested that a more gradual and thoughtful introduction of these technologies would have been preferable. While consumer-facing AI agents attract attention for various issues, enterprise-level applications remain limited in practical impact.

A September survey by Ernst and Young highlighted the risks associated with current deployment strategies. The study found that 36 percent of senior AI executives at companies with annual revenues exceeding one billion dollars have experienced a severe AI incident. These incidents resulted in financial damage, operational disruptions, data loss, or reputational harm. These findings contribute to a growing reluctance among corporations to embrace the AI revolution, even for firms whose own business models depend on promoting AI solutions to clients.