Don’t Be Fooled By AI Washing: 3 Questions To Ask Before Investing

We’re excited to bring Transform 2022 back in person on July 19 and pretty much July 20-28. Join AI and data leaders for insightful conversations and exciting networking opportunities. Register today!

Let’s face it: tech marketing has clouded the waters for investors looking for genuine AI companies. As we all know, not all AI companies have real AI as their core value proposition.

You can divide AI companies into two categories: companies that use AI as a medium, or as an add-on to automate a particular function. And others whose entire raison d’être depends on AI. As an AI practitioner and investor, I am biased. I believe that companies that apply real AI are unfairly grouped with companies that only use AI as a function, often classified under ‘automation’.

We’ve all read this story before. It is reminiscent of the early cloud days, when a branched ecosystem of businesses emerged. On the one hand, we had the companies that were born pre-cloud; on the other hand, we had the “cloud-first” batch. The latter is the equivalent of what we see with AI today – companies born far enough into the AI maturity lifecycle to be able to build companies powered entirely by AI. These are companies that think differently and that unlock real innovation.

Words matter – you can’t just attach ‘AI’ to your solution and call it real AI

Not to introduce more jargon, but I’m going to – because we need clarity in this space. Companies that do real AI are what I call AI native. And if AI is not the reason it exists; it is not AI native. Simply.

Let’s do a quick quiz. Is YouTube AI Proprietary? If yes, then you are disqualified. If no, ding, ding, ding! YouTube only uses AI as a medium to stimulate consumption, but it is not central to the user experience as with TikTok, for example. TikTok decided to make AI work for creators too, with editing tools and improvement suggestions that resonate with the audience they want to reach.

By putting AI at the center of their service, they have created a new innovative video creation category that allows people to change their face, control video elements with facial expressions, etc. That’s what makes many TikTok videos almost impossible not to to be viewed repeatedly.

For another example, take AI-native Insitro, an AI-native company that builds disease models based on large-scale automated experiments with the goal of developing new therapies. By analyzing vast amounts of clinical and genetic data, these models can determine “how a disease works” and help identify new subtypes of disease — ultimately assessing which patients are most likely to respond to a particular therapy. This is only possible thanks to machine learning due to the very large number of confounding factors involved.

Here’s another example that will help crystallize my point. Ask yourself: Why do (EV-native) cars like Tesla and Rivian act so differently than other traditional (gas-first) automakers entering the EV race? The endgame may be the same, but the products are incomparable.

Over the past decade we have seen AI washing take over. In many cases, AI has become a buzzword to mean anything with some form of algorithmic “intelligence” — from boring stuff like insurance and financial management to getting the right ads to people.

Do you want to invest in AI native? Make sure you ask the right questions.

Not sure how to tell an AI native from a non-AI native during a pitch? Lead with these questions and you’ll get the answer pretty quickly:

Q: Why is this real AI? For example, if the answer is something along the lines of “our platform has an automation engine to streamline workflows,” then move on. If you get a vague answer like “it has some data and some logic”, there is something strange. Or if you get a jargon dump that reads, “it’s a database plus a best-in-class advanced query engine”, run. Q: Which built first, the model or the company/solution? If the answer is something along the lines of, “Well, we focused on the user experience and then applied algorithms to streamline the workflow engine,” it’s not AI-native. When people talk about “low in” AI in retrospect, or that we “built an advanced AI engine on it”, then you know that it probably isn’t real AI. Q: Is AI a boardroom conversation at the company? You know AI isn’t the core of the business when… The word AI appears on a slide and everyone is surprised and confused. As the founder says, “the board slides are AI-generated.” If the board meeting starts with a TikTok.

To sum up my argument, I welcome any ethical use of AI. But calling any company that makes any use of AI an “AI company” undermines companies that do deep AI work or AI native products; it is also misleading and potentially risky from an investment perspective.

So the next time you get the idea of using the word AI to describe a business, think about these points, and maybe even ask yourself the above questions. If you eventually realize you’re not AI native, that’s totally okay – just think more about using the word “AI” for the sake of industry clarity. I think we can all agree that we need it.

Luis Ceze is a partner at Madrona Ventures and CEO of OctoML

DataDecision makers

Welcome to the VentureBeat Community!

DataDecisionMakers is where experts, including the technical people who do data work, can share data-related insights and innovation.

If you want to read about the very latest ideas and up-to-date information, best practices and the future of data and data technology, join us at DataDecisionMakers.

You might even consider contributing an article yourself!