Study reveals that 95% of AI generated by MIT fails to generate financial value: What's behind the job cuts?
In the ever-evolving landscape of technology, Generative AI (GenAI) has made its mark, particularly in sectors such as customer support, admin operations, and process outsourcing. However, its impact is more noticeable through attrition rather than mass firings, according to recent observations.
A survey of over 80% of executives in the tech and media sectors predicts lower hiring volumes within the next two years, hinting at a possible shift in the workforce due to AI integration. Yet, the NANDA report, titled "The GenAI Divide: State of AI in Business 2025", reveals a concerning trend: most enterprise use cases of GenAI deliver little to no measurable impact on productivity or profit.
The report further suggests that GenAI is often seen as a corporate status symbol, with a fear of missing out (FOMO) driving adoption. However, AI is increasingly being used as a justification for company restructuring and cost-cutting measures, raising questions about its true purpose and effectiveness.
The integration of GenAI with legacy enterprise workflows is a significant problem, with many companies grappling with this challenge. The learning gap is another hurdle, as generic tools like ChatGPT don't integrate with internal data or learn from institutional knowledge.
Moreover, GenAI models are not designed for predictability, hierarchy, or rule-based data, unlike traditional business systems. This mismatch between AI capabilities and traditional business processes is a key factor in the persistent challenges and delays in successful integration.
The primary responsibility for implementing and integrating GenAI falls on company leadership and line managers. However, many AI pilot projects stagnate, with only about 5% achieving rapid economic impact. Successful integration requires not just technology but also a strong data strategy, governance, AI literacy, and iterative agile methods.
The current state of GenAI is reminiscent of the early days of cloud computing, suggesting that AI may follow a similar trajectory of initial challenges before becoming more seamless. Agentic AI systems, which learn, adapt, and act semi-autonomously, are seen as a potential solution for the current problems faced by GenAI.
However, a puzzling question remains unanswered: why is GenAI causing hiring freezes and layoffs despite its lack of business value? The MIT NANDA report does not directly address this issue.
It's worth noting that most companies have invested their AI budgets heavily in frontend sales and marketing, while back-office automation has been neglected. This imbalance could be a contributing factor to the perceived lack of business value from GenAI.
In conclusion, while GenAI holds promise for the future, its current state is a mixed bag of challenges and opportunities. The question remains whether GenAI is being used to solve meaningful problems or just being chased as a shiny new thing to justify layoffs.
