The AI at Joonko Wasn’t Real
Remember when companies thought it was a good idea to have different kinds of people working for them? The Supreme Court splashed some cold water on that idea last year, and now the companies don’t talk about it as much. But in the heyday of what they called “DEI”, a company called Joonko Diversity promised to use artificial intelligence to make all your DEI dreams come true. According to the SEC and prosecutors in the S.D.N.Y., Joonko offered more than it could deliver. Maybe because its investors included a bunch of venture capital firms that invested millions of dollars, now Joonko’s founder and CEO Ilit Raz is being prosecuted and is probably sad.[i]
Joonko’s Glorious Beginnings
Raz founded Joonko in 2016 to match diverse and underrepresented job candidates with companies that wanted to meet DEI goals. Joonko’s original mission was to use AI and data to identify and solve unconscious gender and racial bias within companies by tapping into their workflow and task management platforms. Later, Joonko sought to use AI to create a pool of “silver medalist” job candidates, i.e., underrepresented candidates that made it to the final stages of a company’s hiring process but didn’t ultimately receive job offers. Joonko would use its technology to match these candidates with job opportunities at its customer companies.
What She Said:
Raz said a lot of things to raise money for Joonko:
About Joonko
In February 2021, Raz sent marketing presentations to investors saying that Joonko had more than 120 companies on its platform and highlighted four Fortune 500 companies in particular. (It never had more than 30, and more than half of those didn’t pay anything to be there. The four Fortune 500 ones never worked with Joonko.)
Another presentation included some fake testimonials:
“Joonko has been immeasurably effective in helping our team attract more diverse candidates, by bringing them directly to our recruiters and keeping track of our hiring progress while advancing these candidates”;
“Diversity is at the heart of our culture and Joonko helped us turn this aspiration to a successful reality”; and
“I love the fact that I don’t need another dashboard and that relevant candidates are simply flowing into our [Applicant Tracking System], saving me so much time on sourcing. Because of Joonko, I was able to get back to what really matters – hiring great talent.”
I bet the AI Raz used to generate these testimonials was real.
Raz also:
sent investors fake customer lists and Joonko bank statements;
claimed in a YouTube video that Joonko was working with over 200 customers, including the Fortune 500 ones; and
listed several Fortune 500 companies as Joonko customers on her LinkedIn profile.
About Joonko’s Customers
It also probably seemed advantageous to have a bunch of “silver medal candidates” in Joonko’s applicant pool. They had some things to say about that! In different forums, they claimed “more than 40,000 candidates in [its] talent pool,” and also 185,000 active monthly candidates (with 4,589 in New York). The SEC’s complaint is a little vague but they may not have had any.
About Joonko’s Revenue
Joonko’s revenue in 2021 or 2022 didn’t get over $100,000, but the story Raz told investors was a lot rosier than that. She told one VC firm that Joonko had hit $1.84 million in revenue in 2021, and another presentation claimed $2.2 million for the same year. She claimed $3 million for 2022. In April 2022, Raz showed one investor a fake bank statement with large incoming deposits from customers and an account balance of $3.7 million that had been as high as $6.8 million. These numbers were not real.
About Joonko’s Platform
Raz said a lot about the Joonko platform’s use of AI and automation, and its ability to connect to customer companies’ applicant tracking systems. She claimed Joonko’s technology was based on “seven different AI algorithms” and then identified them with very data-sciencey language. In an interview, Raz said Joonko’s “proprietary algorithm first uses natural language processing and computer vision to scan public data on the candidates that are referred to us.” Also, Joonko used “machine learning to improve the matching process” which was “automated from end to end.”
The SEC then says with some understatement: “Joonko did not actually use these processes.”
As time went on, Raz’s scheme unraveled in what must have been a very stressful period of outrunning investors and Joonko’s board. After lots of fake contracts, bank statements, and signatures, one of the board members (also an investor) confronted Raz about all of the fake everything and she admitted the scheme.
The Gist
Look, we’re still in the AI-washing phase of AI enforcement. People may be doing clever frauds with it, but they haven’t been caught yet. But I bet it’s just around the corner. Just wait!
Students for Fair Admissions, Inc. v. President and Fellows of Harvard College, 600 U.S. __ (2023)
SEC Charges Founder of AI Hiring Startup Joonko with Fraud, SEC Press Release (June 11, 2024)
SEC v. Ilit Raz, Civ. No. 1:24-cv-04466 (S.D.N.Y., filed June 11, 2024)
[i] The “facts” that follow haven’t been litigated yet and are really just allegations. Who knows if they’re true?