These days, investors, founders, analysts, and tech loyalists can’t stop talking up how artificial intelligence will radically remake the future of work. But a new study suggests that while AI might have some amazing future uses, at the moment, it’s not doing much good to anyone.
In a “State of AI in Business 2025” report, researchers at MIT Media Lab found that despite enterprise investments of as much as $40 billion in generative AI, 95% of organizations have seen no return on their investment so far.
“Just 5% of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable P&L [profit and loss] impact,” the report reads. “This divide does not seem to be driven by model quality or regulation, but seems to be determined by approach.”
Tools like OpenAI’s Chat GPT and Microsoft’s Copilot have been widely tested, with more than 80% of organizations either exploring them or using them in pilot programs. But while these tools boost individual productivity, when it comes to overall P&L, there’s very little measurable impact, the study found.
Enterprise-level AI systems aren’t doing a lot to impress either. While 60% of the enterprise companies that the researchers spoke with said they evaluated these tools, only 20% made it as far as the pilot stage. And just 5% took those to a full production model. A lack of contextual learning, brittle workflows, and misalignment with day-to-day operations were cited as the chief reasons the tools were rejected.
The slop problem
Behavioral researcher BetterUp Labs, in collaboration with the Stanford Social Media Lab, says it has identified one possible reason enterprise companies are rejecting AI: slop. Employees, the two labs say, are using AI tools to create low-effort, passable-looking work. They’ve taken to calling this “workslop”—well-formatted slides, reports, summaries, or code that might seem helpful at first, but that ultimately prove to be incomplete or missing context, shifting the burden of work to someone else.
“Of 1,150 U.S.-based full-time employees across industries, 40% report having received workslop in the last month,” the groups wrote. “Employees who have encountered workslop estimate that an average of 15.4% of the content they receive at work qualifies.”
There is one advantage for workers when it comes to AI not living up to its billing: Job losses, so far, haven’t been as bad as doomsayers have predicted. Should more companies stick with AI integration, however, that could change.
“While most implementations don’t drive head count reduction, organizations that have crossed the GenAI Divide are beginning to see selective workforce impacts in customer support, software engineering, and administrative functions,” the report reads.
Dispelling doom
When it works, AI can cut back-office operational expenses—for example, in administration, finance, and human resources—the MIT researchers found. In addition, it can improve customer retention and sales conversion by using automated outreach and by following up intelligently. Best of all, few of these improvements come with a human cost.
“Early results suggest that learning-capable systems, when targeted at specific processes, can deliver real value, even without major organizational restructuring,” the report reads.
That’s likely a relief to some workers who have heard little but fatalistic scenarios. In July, for instance, Aravind Srinivas, the CEO of Perplexity, warned that recruiters and executive assistants could be made extraneous as AI browsers become more prevalent.
And in May, Anthropic CEO Dario Amodei told Axios that AI could wipe out roughly 50% of all entry-level white-collar jobs within five years, which he said could cause unemployment to spike to between 10% and 20%.
That warning, he said, was aimed at both lawmakers and his peers in the AI world.
“Most of them are unaware that this is about to happen,” Amodei said. “It sounds crazy, and people just don’t believe it. . . . We, as the producers of this technology, have a duty and an obligation to be honest about what is coming.”
The application deadline for Fast Company’s Most Innovative Companies Awards is Friday, October 3, at 11:59 p.m. PT. Apply today.
