The performance of investors in AI-first technology firms serving the defense sector, like Palantir, Primer, and Anduril is encouraging. Anduril for example, has a value of more than $4 billion within four years. Other companies, such as Image Labeling Technologies that develop general-purpose AI-first technology — like image labeling — also receive substantial (undisclosed portions) of their revenues from the defense sector.
AI-first tech companies are often a good investment. Even though they don’t have a purpose to support the defense industry, investors find that their companies eventually help (and sometimes accidentally) other powerful institutions such as municipal agencies, police, and media companies in prosecuting their duties.
Many do great work. DataRobot helps agencies understand COVID spread, HASH runs simulations vaccine distribution, or Lilt makes school communications accessible to parents of immigrant children in U.S. schools districts.
Knowing what’s going on is the first step to taking on responsibility. Startup investors often ignore the fact that AI-based models are complex.
There are some negative examples, however. According to The Washington Post, 16 media partners, technology provided by NSO, an Israeli cyber-intelligence company, was used to hack 37 smartphones of journalists, human rights activists, business executives, and the fiancee murdered Saudi journalist Jamal Khashoggi. According to the report, the phone was on a list with over 50,000 numbers that were based in countries which monitor their citizens. The Israeli company is also known to have employed the service of this firm.
These investors may be confronted by limited partners, governments, and other founders about the power of these technologies, as well as whether they are being applied too widely or enable too many benefits. These questions are of degree but may not be asked when an investor makes an investment.
I have had the pleasure of speaking with many people from many perspectives, including CEOs and founders of large companies as well as current leaders of small businesses. Since publishing “The AI-First Company”, in 2010 and then investing for nearly a decade in these firms, I’ve had the privilege of speaking with politicians and small businesses. Over and over, I have been asking the same question: How can investors make sure that startups they invest in responsibly use AI?
We’ll be honest: Startup investors can easily dismiss such a crucial question with statements like “It’s hard to determine when we invest.” But startups are still a nascent form of what’s to come. AI-first startups have something in common: They are using tools that provide leverage beyond what our bodies, intellects, and time can allow.
AI gives users the ability not only to manipulate heavier objects (robots), but also to analyze more data (analytics). It can even allow them to make predictions about the future. People learn faster when they are able to make accurate predictions and then learn from them. People learn quickly and can take action fast when they do.
These tools can be used for good and for bad, just like any other tool. A rock can be used to make a house, or it can be used as a weapon against someone. Gunpowder can be used to make beautiful fireworks and fire bullets.
AI-based computer vision models are substantially similar and can be used for figuring out moves of terrorist groups or dance groups. Ai-powered drones are able to aim at us with a camera while we do ski jumps. They can also point a gun at our faces.
This article explains the politics, statistics and basics of investing responsibly in AI-first businesses.
AI-first company investors and members of their boards must assume at least some responsibility for decisions made by the companies they invest in.
Whether they want to or not, investors influence founders. Investors are constantly questioned by founders about which products they should build and how to reach them. This is to improve and learn, as well as increase their chance of success. This is partly why they keep investors informed and engaged as it may prove to be valuable capital.
Publiated at Wed 15 Sep 2021 21.25:25 +0000