When ChatGPT was released to the general public, it instantly became the number one subject and remains a constant for companies. Consider senior executives mentioned the terms AI or artificial intelligence an average of 3.7 times a call with analysts in the second quarter or more than double the year before. AI remains a topic which every company says they will use, but is it effective?

In an article by Sam Sivarajan, it is useful to understand the limits of AI, at least at the moment. A recent Harvard Business Review article puts it: Artificial intelligences are prediction machines. They can tell you the probability it will rain today, but not whether you should bring an umbrella. The umbrella decision requires more than prediction.

The decision requires a judgement, which reflects individual preferences and experiences. When the forecast is 10% rain, some will take an umbrella, and others will not, it comes to personal preferences.

For your investments, you see individual risk tolerances all the time. Some will buy high risk low cost stocks, others will buy dividend paying stocks. The probability of a loss (or gain) from those investments is the same for all investors. Some investors have a preference and tolerance for higher probabilities of loss than others. This is why determining an investor’s risk tolerance is not a straightforward exercise.

In the Major League Baseball World Series won by the Texas Rangers, one of the stars of the Rangers was Jordan Montgomery. Last year he played for the New York Yankees, the Yankees traded him because they did not trust him to win big postseason games.

There are limits to the prediction machine. It does not factor in individual human preferences or experiences. Nor does it account for learning, adapting or adjusting on the fly. In the baseball game the batter and pitcher are not the same in the latter innings as they were in the earlier ones. The prediction machine cannot account for that, yet.

These limits of data and prediction machine can have costly implications for companies and investors. In February, the tech based real estate company Zillow Group Inc, set up its AI to value homes and make cash bids. By November the company stopped doing it because the homes it bought could not be sold for higher prices. The company had to do a $304 million inventory write down. The stock fell and 25% of staff were laid off.

The point is AI will be and is very valuable to analyze reams of data and provide empirically testable conclusions, which save valuable time. But humans should be involved in making the final decisions. because context is important.

Linking to dividend paying stocks, there are some industries and some companies that using AI to value the company should be easy as clockwork, unless the company is doing something illegal. But more stable companies, ensuring they have consistent revenues, their margins have not fallen and they are profitable can be relatively easy. It is the growth companies that an expectation of growth is needed that requires judgement. Generally as an investor you expect the company to be using AI and making better decisions to ensure the company remains competitive and profitable.

There are more questions than answers, till the next time – to raising questions.