Prediction: These 16 Words From IBM Will Prove to Be Prophetic About Artificial Intelligence's Future

Opportunities related to artificial intelligence (AI) seem endless, particularly in deploying agents. There's potential for AI agents to significantly overhaul processes and take on multi-step tasks that humans handle today, which is why there's growing concern about AI-fueled job losses.

But with that opportunity also comes risk, including the danger of costly AI errors that could cost a company millions of dollars or more. There's one intriguing phrase from an old International Business Machines (IBM) training manual that I believe is incredibly relevant today and will likely prove prophetic in the future.

Image source: Getty Images.

IBM's old training manual identifies a simple but often overlooked problem with AI

Although generative AI is new, computers have been revolutionizing businesses for decades, with IBM often at the forefront of that transformation. And even back in 1979, its training manual highlighted a huge problem with relying on computers to do work, which also pertains to AI today: a lack of accountability.

"A computer can never be held accountable, therefore a computer must never make a management decision."

If AI makes a mistake, who is accountable and responsible? There isn't a clear answer, yet it's something I wonder about whenever I see new AI developments in which an AI is going to make purchasing or investment decisions on behalf of a person. While it may seem like it may help simplify someone's life, it can also significantly increase the risk. AI chatbots have shown they can be manipulated, suggesting agents may become highly vulnerable to scammers.

Are AI stocks overhyped?

Tech giant Nvidia (NASDAQ: NVDA) has been generating strong growth in recent years, driven by AI and the investments companies have been making in next-gen technologies. Expectations of continued growth are why Nvidia, with a market cap of more than $5 trillion, can still be seen as a relatively cheap buy; it trades at a forward price-to-earnings (P/E) multiple of only 24 (based on analyst expectations). That's not too pricey for a top growth stock, given that the S&P 500 average forward P/E is around 22.

The danger, however, is assuming there won't be challenges along the way, particularly as AI assumes greater responsibility for companies. I believe that as that happens, companies will also more carefully reevaluate the risks related to AI and ultimately return to the same conclusion that IBM reached decades ago: some tasks aren't suitable for AI due to a lack of accountability and responsibility. If that happens and businesses scale back on AI expenditures, the entire thesis of investing in Nvidia and other AI stocks could quickly unravel, as their valuations may suddenly look less attractive.

This is why I believe it's crucial for investors to be cautious with AI stocks: while many have performed exceptionally well in recent years, they are by no means risk-free investments.

David Jagielski, CPA has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends International Business Machines and Nvidia. The Motley Fool has a disclosure policy.