Amazon axes ‘sexist’ AI recruiting tool
Program was reportedly biased against women

Amazon has reportedly scrapped a recruitment tool powered by artificial intelligence (AI) that showed a clear bias against women.
Company insiders told Reuters that the program had been in development since 2014 as a means of streamlining the recruitment process.
“Everyone wanted this holy grail”, one of the sources said. The system was designed to sift through 100 job applications and “spit out” the top five results.
Subscribe to The Week
Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

Sign up for The Week's Free Newsletters
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
The machine gave each applicant a rating out of five, “much like shoppers rate products on Amazon”, says the news site.
The AI was trained to use a method known as machine learning. This allowed researchers to feed the program old CVs from the past ten years to use as reference material so the AI could accurately grade real-life job applications, the BBC says.
But the data submitted came primarily from former male applicants, meaning the “sexist” AI taught itself that “male candidates were preferable”, the broadcaster reports.
The bias was discovered in 2015 when recruiters found that the AI system wasn’t rating applicants in a gender-neutral way. The program was penalising CVs with the word “women” in them, putting female applicants who had studied at women’s colleges, for example, or written “women’s chess club captain” in their list of achievements at a disadvantage.
Programmers subsequently amended the system to remove the bias but this was not a complete success, according to Fortune. The AI was abandoned last year.
Amazon declined to comment on the technology, but said the tool “was never used by Amazon recruiters to evaluate candidates.”
Has the AI hindered female job applicants?
Amazon insiders insist it hasn’t.
Speaking to The Sun, Amazon sources said the retail giant “never based recruitment decisions on the tool's ratings alone”, but recruiters did examine the AI’s recommendations when sifting through applications.
This isn’t the first time an AI programme has been shown to favour men over women.
Facebook came under fire last month after it emerged that AI systems had been preventing women from seeing certain job advertisements, the newspaper says.
Sign up for Today's Best Articles in your inbox
A free daily email with the biggest news stories of the day – and the best features from TheWeek.com
-
Selling sex: why investors are wary of OnlyFans despite record profits
In The Spotlight The platform that revolutionised pornography is for sale – but its value is limited unless it can diversify
-
Garsington Opera opens its summer festival with two 'very different productions'
The Week Recommends A 'fabulous' new staging of Tchaikovsky's The Queen of Spades and Donizetti's fake-love-potion comedy L'elisir d'amore
-
The Rehearsal series two: Nathan Fielder's docu-comedy is 'laugh-out-loud funny'
The Week Recommends Television's 'great illusionist' has turned his attention to commercial airline safety
-
Google's new AI Mode feature hints at the next era of search
In the Spotlight The search giant is going all in on AI, much to the chagrin of the rest of the web
-
How the AI takeover might affect women more than men
The Explainer The tech boom is a blow to gender equality
-
Did you get a call from a government official? It might be an AI scam.
The Explainer Hackers may be using AI to impersonate senior government officers, said the FBI
-
What Elon Musk's Grok AI controversy reveals about chatbots
In the Spotlight The spread of misinformation is a reminder of how imperfect chatbots really are
-
Is Apple breaking up with Google?
Today's Big Question Google is the default search engine in the Safari browser. The emergence of artificial intelligence could change that.
-
Inside the FDA's plans to embrace AI agencywide
In the Spotlight Rumors are swirling about a bespoke AI chatbot being developed for the FDA by OpenAI
-
Digital consent: Law targets deepfake and revenge porn
Feature The Senate has passed a new bill that will make it a crime to share explicit AI-generated images of minors and adults without consent
-
AI hallucinations are getting worse
In the Spotlight And no one knows why it is happening