By now, you've been hearing it just about everywhere in the media...
AI is changing the way businesses and humans work, live, and invest.
Looking ahead, the technology will continue to shape the world in ways we can't predict now.
However, we can't get excited about every AI startup that promises to innovate in new ways...
In fact, there's a concerning new trend that four MIT students discussed in a recent report.
It helps show why investors need to be cautious about throwing money at every AI-related business...
AI at Work Isn't What You'd Expect Right Now
I'm talking about an MIT report called "The GenAI Divide: State of AI in Business 2025." It covers the results of interviews, surveys, and analysis of more than 300 publicly disclosed AI initiatives.
And a big takeaway from the report might sound surprising...
Ninety-five percent of organizations that invested in "GenAI" tools made no returns on the investments.
You might have seen the term "GenAI" before. It's an abbreviation for "generative AI." And it refers to AI that creates new and unique data from the data sets that the technology trained on.
The MIT report analyzes data from GenAI and general-purpose large language models ("LLM") – AI tools that users can "talk" to for information. Models like ChatGPT and Claude are examples of LLMs.
Overall, the report states that organizations have higher rates of successful use with LLMs than with GenAI tools.
Users were enthusiastic about using GenAI programs for themselves at first. But these programs failed to deliver reliable, usable results.
According to the report, these GenAI tools also had poor user experiences. And users often couldn't trust the product it spat out.
Users were also unwilling to change their routines to add GenAI into their workflow.
I don't blame them, though. I wouldn't want to change a functional routine to a less-efficient one.
However, the MIT report states that employees saw LLMs in a better light...
In fact, more than 80% of users said an LLM gave better answers than task-specific GenAI.
However, we can't view this as a total vote of confidence in LLMs...
According to the report, only 10% of enterprise-level LLM users said they use it for complex, multiweek projects. A larger 70% said they only use LLMs for quick tasks.
These users felt that LLMs didn't respond to their feedback fast enough. LLMs learn from consuming data in huge volumes. Enterprise users can't specialize the LLMs quickly enough to turn profits.
The report also stated that every prompt also needed large amounts of context to create usable results. The time spent providing the LLM context and editing the results can offset the time savings of using AI.
However, the report says that use of AI has improved a few marketing metrics...
Companies saw a 10% improvement in customer retention when using AI.
And thanks to AI, lead qualification speed was 40% faster. Lead qualification is the process of evaluating the chance that a customer will purchase a particular product.
But the report says that organizations got better use from AI in "back office" roles.
Organizations saw faster document-processing speeds with AI. And AI reduced the fees and costs associated with outsourcing work.
Don't Blindly Pile Into Every AI-Related Stock
Earlier this month, I noted that the development of AI is obviously creating big winners in the tech industry. But I also cautioned about falling into the trap of "tunnel vision."
Of course, investors would miss huge gains by ignoring the AI boom altogether. However, as investors, we need to be wary of every tech startup with a flashy AI product...
There's a difference between rolling out an AI product that promises big results... and the product actually achieving those results.
As we've said many times here at the Chaikin PowerFeed, it's a similar case with other booms in game-changing technology. Amid the big winners, there will also be plenty of losers.
We must pick our AI investments carefully. It's not as simple as expecting "easy money" from AI stocks.
Good investing,
Ethan Goldman
