Strategies to defeat AI bots

The writer is a professor of finance at New York University’s Stern School of Business

A few months ago, I was teaching my classes at Stern when my friend Vasant Dhar, who teaches a range of courses from machine learning to data science, called me about the Damodaran Bot.

This is an AI creation that has read everything I’ve ever written, watched every webcast I’ve ever posted, and reviewed every rating I’ve ever posted. He told me that the Bot was ready for a trial and ready to value companies. These ratings could then be measured against the ratings made by the top students in my class.

The contest results are still tabulated, not sure what results I’d like to see. If the AI ​​values ​​companies as well, or better, than I do, that’s a strong signal that I’m facing attrition. If it goes that badly, that would be a reflection of my failure as a teacher.

AI is the union of two forces—increasing (and cheaper) computing power and the accumulation of data, both quantitative and qualitative. As an AI novice, there are three dimensions where I see it having an advantage over humans: on mechanical/formulaic work as opposed to intuitive; in disciplines based on rules rather than principles; and on tasks with there is an objective response rather than subjective judgments. Boiling it down to the personal, the threat to your job or profession from AI will be greater if your work is largely mechanical, rule-based, and objective, and less if it is intuitive, principled, and open to judgment .

While AI in its current form may not be able to replace you at work, it will improve over time and learn more by watching what you do. So what can you do to make it harder to be outsourced by machines or replaced by AI? I have four thoughts.

First, in a world of tunnel-visioned, siloed specialists, AI will empower generalists, comfortable across disciplines, who can see the big picture.

Second, in investing and valuation, if your valuation technique has become primarily financial modeling with extrapolation from past data, AI can do it faster and with much less error than you can. If, however, your assessments are built around a business story, enriched with soft data, AI will have a harder time replicating what you’re doing.

Third, we fall victim to the curse of “Google Search” where when faced with a question, we rush to look for the answer online rather than trying to find the answer. Although benign if you’re looking for answers to trivia, it can be malignant when used to answer questions we should be reasoning out for ourselves. This reasoning may take longer and lead you to wrong answers, but it’s a learned skill and one we risk losing if we let it languish.

Fourth, an empty mind may be the devil’s workshop, but it is also the birthplace of creativity. The ability to connect seemingly unrelated facts and have “Aha” moments is unique to humans, and AI will struggle to do the same.

If you were a conspiracy theorist, you could tell a story of tech companies conspiring to deliver us products, often free and convenient to use, that make us more specialized, more one-dimensional, and less reason-based, while we fill our free time as a precursor to AI being unleashed as a weapon on us.

Since my life’s work is in the public domain and there is a bot with my name on it, my AI threat is here. The AI ​​threat to you may not be as imminent, but as you think about the answers, there are three strategies you can try. The first is to be so secretive about what you’re doing that a bot won’t be able to track you. The caveat though could be that your actions can reveal your work process and the AI ​​can reverse engineer what you’re doing.

The second is to seek system protection, from regulators and the law, against AI disruption. So even if AI can replace humans in valuation, I’ll bet that courts and accounting rule-makers will be convinced that the only acceptable valuations can come from human valuers. The third is to build a “moat” – strategic defenses that will make it harder for AI to replace you at work. This, however, will require an honest assessment of what you bring to the job.

If you think I’m overreacting to the threat of AI, I’m suggesting that whether the threat is real or imagined, the costs of being real are so significant that we need to act as it is, and act now, because those actions. it will make us all better at what we do, even if it turns out to be imaginary.