The AI myth that has propelled Wall Street

Now CEO's will call anything "AI" just to get a higher multiple for their stock

Not long ago, no one but techies had ever heard of ChatGPT. Artificial intelligence was the subject of fiction novels imagining a future where computers were “sentient”, robots controlled society, and humans were redundant.

Along came ChatGPT and it immediately captured the imagination of the investing public, driving chip stocks like NVidia and software giants like Microsoft to unheard of multiples of earnings and cash flow and prices investors had only dreamed about.

Immediately, every company became an “AI” company. Large language models (LLM’s) became the buzzword of the day, and a plethora of consulting firms emerged within weeks claiming to be able to guide business leaders on the effective use of AI. In many ways it is laughable.

ChatCPT is little more than a Google Search on steroids with a clever algorithm to provide plain language responses to queries, and the use of Microsoft’s brilliant image creator allowed non-artists like me to generate quite satisfactory pieces of art with simple instructions to the similarly brilliant interface. Here’s an example of a piece of art I created in five minutes using Image Creator. Before that time, my attempts at art were akin to stick figures and I couldn’t paint my living room without making a mess of it.

Image Creator and ChatGPT are terrific advances and demonstrate that AI exists, is quite advanced and has utility. But why is it that every public company everywhere now seems to claim it is using AI to run elements of its operations? I think a firm using little more than an old HP80 calculator would now claim that to be AI, the benefits on Wall Street being too hard to ignore.

Tesla makes cars, but Elon Musk says Tesla is an AI/Robotics company. Musk was one of the founders of Open AI and has the credentials to make such a claim even if the use of AI in Tesla proves to be quite limited.

Pfizer is a drug company but overnight it claims its use of AI will make drug advances speed up.

Kraft Heinz makes packaged food but now claims generative AI tools will expand the company’s “data driven insights to employees in real time”. What did they do before - eschew data or delay giving insights they found to employees?

Like all manias, the race to claim effective use of AI to propel share prices to higher levels may do just that, making CEO stock options more valuable but lowering returns to investors who will have to pay more to buy into the same earnings stream they could previously acquire at a lower price.

The message for investors is an old one - if it sounds too good to be true, it is too good to be true. Kraft Dinner will not taste better because Kraft Heinz employees get data driven insights in real time and Pfizer couldn’t have produced mRNA any faster using magic (and many people who have been harmed by Pfizer’s mRNA might have preferred Pfizer to use the conventional longer-term approach to testing to surface the often lethal cardiac side-effects that probably claimed as many lives as Pfizer claims the mRNA vaccine saved).

NVidia will benefit from the demand for more of its chips and competitors will design and develop competing chips and the bloom will come off the rose within a few years. Microsoft will benefit from growth in Azure and other cloud based technologies will advance to ensure the benefit of cloud based AI tools is not solely a Microsoft product. Investors will pick and choose and the best outcomes will result from choosing the few companies that actually have measurable and important advances from the use of large language models and ignoring the ones that do the same old thing they have always done but try to persuade you that pickles and peanut butter are better products becuase of “data driven insights” given employees in real time.