How we define and identify responsible AI enabling companies.

There are a lot of companies out there that profit on the buzzword "responsible." We make sure they don't make it into EAIDB.

Our verification process follows a set of very simple, very transparent rules. Even when there are exceptions, when we make them we try to be as clear as possible on the reasons why.

Verification Procedure

  1. There are a couple of pre-requisites that we impose to maintain the consistency of the database.
    • Founded post-2015. We focus on early-stage companies because this is where direct innovation happens. The cutting-edge isn't seen in larger companies until years after it emerges. As a whole, this industry isn't much older than 2015 anyway. There are a few exceptions in EAIDB, but these are primarily because they are relatively quiet companies that are still early in their product/service journey despite being founded so long ago.
    • Series C or younger. For the same reasoning as above, matured companies are still considered in reports and analysis, but just aren't within scope to consider "cutting-edge."
    • The company's main line of business must fit within what we define as "responsible-enabling." See below for a more in-depth description of what this means.
    • Active. We don't consider companies, open source repositories, or firms that have not provided an update on social media, blogs, etc. within the last year.
  2. We look for evidence of "responsible-enabling." When a company is found (either through their own submission or when we find them), the first resource we look at is their website. Some signs of "responsible activities" might include the following.
    • Active research teams working on subjects within responsible AI like explainability, fairness, etc.
    • Products or services that look like they directly fit the descriptions of one of our eight categories (outlined below).
    • Corporate social media posts, blogs, etc. that frequently discuss topics in responsible AI enablement.
    • Board of directors and founders' histories and personal social media posts for topics in responsible AI enablement.
    • If we find sufficient evidence of these, we consider the company "indirectly verified." This means we haven't had the chance to see a demo or have a 1:1 conversation, but we're fairly certain the company fits the bill.

  3. We grab time on the founders' calendar and discuss the specifics and technical details (and sometimes get a demo). This is how startups become "directly verified" on EAIDB.
    • We've directly verified about 45% of the full database. We're a small team and we're constantly working to increase this number!

Ecosystem Categories

We've identified eight different categories of responsible AI enabling startups. These are not fixed and are constantly iterating depending on what we see in the marketplace. In the past, we used five categories: Data for AI, Targeted Solutions and Horizontal Tech., MLOps and Monitoring, Consulting, and Open Source.