The author is a professor of finance at New York University’s Stern School of Management.
A few months ago, while I was teaching a class at Stern, I got a call from my friend Vasant Dhar, who teaches classes ranging from machine learning to data science, about the Damodaran bot.
This is the work of an AI, which has read everything I’ve ever written, watched every webcast I’ve posted, and reviewed every rating I’ve published. He told me that the bot is ready for pilots and ready to evaluate companies. These ratings can be compared to the ratings given by the best students in the class.
The results of the contest are still being compiled, but we don’t know what the results will be. If AI values companies as much or better than I do, that’s a strong signal that I’m facing obsolescence. If it’s bad, then I’ve failed as a teacher.
AI is the combination of two forces: increasing computing power (and lowering costs) and accumulating quantitative and qualitative data. As a newbie to AI, there are three aspects where I think AI is better than humans. In a field that is rule-based rather than principle-based. And tasks have objective answers, not subjective judgments. To put this into a personal context, the threat posed by AI to your job or profession is greater if your job is primarily mechanical, rules-based, and objective; based, and not much when there is room for judgment.
AI may not be able to replace your job in its current form, but it will improve over time and learn more by observing your behavior. So, what can we do to make it more difficult to outsource to machines and replace by AI?I have four ideas.
First, in a world where experts operate in silos and exhibit tunnel vision, AI will empower generalists who are comfortable looking beyond their areas of expertise to see the big picture.
Second, in investments and valuations, where financial modeling by extrapolating historical data is the primary valuation method, AI can do so more quickly and with far fewer errors than humans. But if your reputation is built around your business story and enriched with soft data, it will be harder for AI to replicate your actions.
Third, we are victims of the curse of “Google Search.” When faced with a question, instead of trying to find the answer, we immediately search for the answer online. It’s benign when looking for trivia answers, but can be malignant when used to answer questions that require you to reason out the answer yourself. That reasoning takes time and may lead you to the wrong answer, but it’s a learned skill that you risk losing if you slack off.
Fourth, an empty mind may be the devil’s factory, but it is also the birthplace of creativity. The ability to connect seemingly unrelated facts and experience “aha” moments is unique to humans, and it will be difficult for AI to do the same.
If you’re a conspiracy theorist, you know that technology companies are colluding to provide us with products that are often free and convenient to use, filling our free time while making us more specialized and more one-dimensional. You can weave a story that makes it more objective and de-emphasizes reason. This is a sign that AI will be unleashed on us as a weapon.
My life’s work is in the public domain and there are bots that include my name, so my AI threat is here. While the threat from AI may not be immediate, there are three strategies you can try when considering a response. The first is to keep your actions secret so bots can’t track you. However, the caveat is that your actions may reveal your work processes, and AI may be able to reverse engineer your actions.
Second, we need regulators and laws to protect our systems from AI disruption. Therefore, even if AI could replace humans in appraisal tasks, I would wager that courts and accounting rule makers would be persuaded that acceptable appraisals can only be provided by human appraisers. The third step is to build a “moat.” This is a strategic defense that makes it more difficult for AI to replace your job. However, this requires an honest assessment of what you bring to the job.
If you think I’m overreacting to the threat of AI, whether the threat is real or imagined, the cost of the threat being real is so significant that I think we should all do so. I suggest that we have an obligation to act and to act now. Even if it turns out to be imaginary, it will make all of us work better.
