Things That Matter

Language Learning App Duolingo Has Been Teaching Phrases Like ‘He has to be detained right now’ In Spanish

It seems like everyday we uncover stories of discrimination and racism towards communities of color. This week the story stems from the most unthinkable source. One of the world’s most popular language-learning apps, Duolingo, served up two very problematic sentences to a user. His girlfriend took to Twitter to let everyone know.

Needless to say that Twitter was not having it.

Two xenophobic statements popped up on a user’s language-learning app and this is why they’re hurtful.

The fact that sentences like: “Are they legal?” and “He has to be detained right now.” can appear on Duolingo has us floored. They may not be explicitly referring to immigration issues, but the choice of words used is loaded with anti-immigration rhetoric. The word “illegal” carries major stigma in a time of pervasive and systematic civil rights violations against immigrants in the United States. Not to mention the weight that a phrase like “He has to be detained right now” carries in a country where at least 2,654 migrant children – and perhaps thousands more – were taken from their parents and held in government custody while their parents were criminally prosecuted for crossing the border unlawfully.

Duolingo was quick to respond to the backlash on Twitter, arguing that the sentence “Are they legal?” did not refer to people but rather questioned the legality of objects namely “firearms.”

In Spanish, “ellos” refers to people, the sentence is clearly about humans.

However, as several Spanish-speaking users noted on the social platform, the structure of the sentence in Spanish, “¿Son ellos legales?”, leaves no room for doubt on whether or not it’s referring to things as opposed to people. If we go back to Spanish grammar 101, Duolingo should know that the use of the plural personal pronoun “ellos” (they) instead of the demonstratives: “eso/esas” or “aquello” (those or that), to point to things, like guns or drugs, make it clear that this sentence is questioning the legality of a human subject. In simpler words, in Spanish “ellos”: refers to people, “esos” to things.

As one twitter user pointed out, “the exercise is not only grammatically incorrect, it leaves it open to a dangerous interpretation.”

Duolingo crowdsources its content from volunteers, and no scarcely anybody moderates what makes it onto the app.

What’s more, it looks like Duolingo’s been slipping on their content for a while now. According to an article published last year on the apps crowdsourcing’s strategy; “The startup has built one of the world’s most popular language-learning apps while only hiring a handful of translators.” Each day the platform serves up millions of sentences, “almost all of them created by its 300 or so volunteers.”

According to Quartz, “more than half” of Duolingo’s employees work in engineering, while “just three people manage the volunteer community”. The app makes this system work by appointing “volunteer moderators”, who apply through the website to oversee content in each language. The company claims that volunteers “are vetted for language skills, and then trained on subjects from pedagogy to guidance dealing with gender, diversity, and cultural sensitivity issues.”

As recent stories of neglectful treatment of migrants, civil rights violations, discriminatory behavior and racial profiling, have become the norm in the media—result of an anti-immigration administration—it’s troubling that this intolerant rhetoric is now visible in apps that are meant for ‘educational’ purposes and that claim to have ‘guidance on diversity and cultural sensitivity issues’.

It’s no coincidence that ever since Donald Trump became President of the United States, there has been a spike in hate speech and crime—and the numbers only continue to climb. A survey of Mexicans recently deported from the United States found that the number of people who reported experiencing verbal abuse or physical assault during their time in the U.S. increased by 47% between 2016 and 2017.

Founder of Duolingo, Luis von Ahn, a Guatemalan immigrant himself, chimed in to the conversation on Twitter, commenting that the two statements “didn’t appear next to eachother” and “were taken out of context”. He assured users that the statements were removed from the app to “avoid confusion”.

If no one is policing the sentences that go up on Duolingo, are children safe to use the app?

But what context would ever make those two sentences necessary in an educational app? They are grammatically incorrect and hurtful. No human is ever illegal. The very object of hate speech is to deprive people of the assurance that society regards them as people of equal dignity. Why is this instance of ‘indirect’ hate speech so important? Precisely because the public conception of immigrants and communities of color, specifically Latinx communities, is constantly under fire, and further feeds into a climate of fear.

As one twitter user put it, it makes you think twice before letting your 8th grader use the app to learn a new language. If xenophobic statements like these can pop up at any time, with no real policing of incorrect, racist or straight-up inappropriate content, then what’s the point of Duolingo even having a ‘content policy’ at all? As a Guatemalan, von Ahn should do better.

Notice any needed corrections? Please email us at corrections@wearemitu.com

Twitter’s AIs Prefer Ted Cruz With Boobs And White Skin Over Black

Things That Matter

Twitter’s AIs Prefer Ted Cruz With Boobs And White Skin Over Black

Ever notice how on some social platforms like Twitter or Instagram that you yourself are mysteriously unable to crop your display images on your own? That’s because Twitter prefers to let their algorithms make the decision. Over the weekend users on Twitter discovered the surprising dangers of letting algorithms crop your own images.

Education tech researcher Colin Madland drew attention to the issue while speaking out about how the video-calling program Zoom, often crops the head out of his black person coworker while on calls.

It didn’t take long for Madland and other users to discover that Twitter’s AIs use discriminatory equations to prioritize certain faces as well. In short, the social platform’s AIs prefer white faces over Black ones.

In response to the discoveries, a Twitter spokesperson acknowledged that the company was looking into the issue “Our team did test for bias before shipping the model and did not find evidence of racial or gender bias in our testing. But it’s clear from these examples that we’ve got more analysis to do. We’re looking into this and will continue to share what we learn and what actions we take,” they stated.

Of course, Madland’s discovery is nothing new. In 2019, test results from the National Institute of Standards and Technology revealed that some of the strongest algorithms online were much more likely to confuse the faces of Black women than those of white women, or Black or white men. “The NIST test challenged algorithms to verify that two photos showed the same face, similar to how a border agent would check passports,” Wired points out. “At sensitivity settings where Idemia’s algorithms falsely matched different white women’s faces at a rate of one in 10,000, it falsely matched black women’s faces about once in 1,000—10 times more frequently. A one in 10,000 false match rate is often used to evaluate facial recognition systems.”

Still, it didn’t take long for users on the platform to ask what other physical preferences Twitter has.

Turns out the AIs prefer Ted Cruz with large anime breasts over a normal-looking Ted Cruz.

(To better understand this Tweet, click the link above)

The user who tested the image of Cruz, found that Twitter’s algorithm on the back end selected what part of the picture it would showcase in the preview and ultimately chose both images of Cruz with a large anime chest.

It’s nothing new that Twitter has its massive problems.

For a platform that so controls and oversees so much of what we consume and how we now operate, it’s scary to know how Twitter chooses to display people with different skin tones. The round of jokes and Twitter experiments by users has only revived concerns on how “learning” computer algorithms fuel real-world biases like racism and sexism.

Notice any needed corrections? Please email us at corrections@wearemitu.com

Alexandria Ocasio-Cortez’s Response To GOP House Candidate’s ‘Dumb Blonde’ Joke Will Leave You Breathless

Things That Matter

Alexandria Ocasio-Cortez’s Response To GOP House Candidate’s ‘Dumb Blonde’ Joke Will Leave You Breathless

BRITTANY GREESON / GETTY

Dumb blonde jokes. They’re overwrought, trite, and pretty outdated. So it’s no surprise that one that came straight from the mouth of a GOP House candidate and directed at Congresswoman Alexandria Ocasio-Cortez didn’t go over so well.

GOP House candidate Marjorie Taylor Greene attempted a joke at Ocasio-Cortez’s expense on Twitter Monday and it failed miserably.

Greene, whose Georgia campaign is being supported by Donald Trump, attempted to hit at AOC’s intelligence on Monday in a tweet.

“As a blonde woman, I would like to take a moment to thank Congresswoman @AOC. She has single handily [sic] put an end to all ‘dumb blonde’ jokes. Blondes everywhere appreciate your service and your sacrifice!” Greene tweeted.

In response, Ocasio-Cortez retweeted Greene, writing, “Don’t worry Mrs. Greene, I completely understand why you need to swing + miss at my intellect to make yourself feel better. You seem to have some trouble spelling your own insults correctly. Next time try ‘single-handedly,’ it’ll work better.”

She signed off her tweet writing “Good luck writing legislation!”

It’s not the first time Greene has come for AOC and failed.

Greene attempted and failed to get under AOC’s skin earlier this month.

Facebook.com

In September, the candidate shared a photo of herself holding a rifle next to images of AOC and other Democratic lawmakers Ilhan Omar and Rashida Tlaib to her Facebook page.

She captioned the post “Squad’s worst nightmare.” It was soon removed by Facebook who cited violations of its policies.

It didn’t take long for AOC’s supporters to strike back at Green to defend the congresswoman.

“Those who are jealous and envious of others typical attack those whom they envy because they need to feel important and try to gain some attention for themselves,” one user commented in the thread. “You are where you are @AOC because of your work and dedication. Mrs Greene knows she can’t compete so she attacks.”

According to People, “Greene has a track record of embracing false stories publicized by QAnon, a conspiracy theory-fueled group which alleges there’s a group of Democratic pedophiles operating around the country.”

Notice any needed corrections? Please email us at corrections@wearemitu.com