AN MIT STUDY PROVIDES INSIGHT INTO HOW TO BEST USE AI
Last week, MIT’s NANDA initiative (Networked-Agents and Decentralized AI) published a study that found 95% of organizations investing in Generative AI are getting zero return on their investment.
The report led to panic among investors, with shares of many tech companies falling, and speculation about a potential AI bubble bursting. Looking past speculation, what did the report actually have to say?
The primary cause for generative AI implementation failing is not due to AI, but instead a “learning gap.” The report points to the failure of generic tools like ChatGPT, which excel for individuals but fall short in enterprise because they don’t adapt to or learn from workflows.
Data from the report also shows that there is a significant misalignment in where AI is being put to use, usually in sales and marketing. MIT found that the greatest ROI is in back-office automation.
Another important factor of the “learning gap” is implementation. Purchasing already made AI tools from specialized vendors leads to a 67% success rate, while attempting to build an AI tool internally increases the chance of failure.
Much like the dot-com bubble, we are now seeing the dangers of investing based on hype or popular speculation. While the explosion of companies investing in and implementing AI has led to a few success stories, Footprints emphasizes the importance of steady growth over time, rather than following potentially short-lived trends.
