Ctrl + N
Rep. Alexandria Ocasio-Cortez (D-N.Y.)’s social media savvy and clout could uniquely position her to take on a complex problem that has so far eluded her more senior peers — potential bias in algorithms used for policing.
Over the weekend, she sounded the alarm about new research that found the facial recognition software Amazon is selling to law enforcement falls short on tests for accuracy and bias. As my colleague Drew Harwell wrote, researchers said Amazon’s algorithms misidentified the gender of darker-skinned women in about 30 percent of their tests. (Amazon founder and chief executive Jeffrey P. Bezos owns The Washington Post.)
The problem stems from the sets of photos the algorithms were trained on — which skew heavily toward white men, the researchers said. Ocasio-Cortez tweeted a link to an article about the research on Sunday, saying that “machines are reflections of their creators.”
When you don’t address human bias, that bias gets automated.— Alexandria Ocasio-Cortez (@AOC) January 27, 2019
Machines are reflections of their creators, which means they are flawed, & we should be mindful of that.
It’s one good reason why diversity isn’t just “nice,” it’s a safeguard against trends like this ⬇️ https://t.co/NcOivu5ejR
Ensuring the algorithms that can help identify criminal suspects are free from bias is a new issue -- and may not be one on most Americans' radars just yet. But Ocasio-Cortez, 29, is already using her intimate knowledge of social media tools, from Twitter to Instagram Stories, to spark a national debate about this challenging and cutting-edge policy issue.
Though she's a rookie in Congress, Ocasio-Cortez commands an impressive online platform: She has amassed a larger social media following than most other members of her party. She has 2.68 million Twitter followers, surpassing House speaker Nancy Pelosi's 2.14 million.
And as a new member of the House Oversight Committee, she also has more tools than a 280-character missive. She is well positioned to examine how the government is increasingly procuring facial recognition technology and whether it is testing for or implementing safeguards for bias. She could also press companies such as Amazon for more information on their technology.
In regards to this particular study, Amazon has disputed the researchers’ findings, saying the algorithms tested by the researchers are different from the algorithms powering the programs that the FBI and police departments in Washington and Florida are using.
But even before the new research was published, with diversity in the spotlight on Martin Luther King Jr. Day last week, Ocasio-Cortez took the opportunity to highlight bias in algorithms during an interview with Ta-Nehisi Coates. “Algorithms are still made by human beings, and those algorithms are still pegged to basic human assumptions,” she said.
Her comments were later slammed by conservatives, including in a report fromThe Daily Wire. The controversy sparked several media outlets to do stories about bias in algorithms, largely saying Ocasio-Cortez is right.
Here's how MIT Technology Review responded to The Daily Wire reporter Ryan Saavedra's Tweet:
She also received support from AI researchers and scientists.
Kate Crawford, a co-founder of New York University's AI Now Institute said:
Time for a collective eyeroll from all of us who actually research this issue. @AOC was talking about facial recognition systems. They're proven to have bias issues. But this is how it gets reported. https://t.co/skVkyRYxZN— Kate Crawford (@katecrawford) January 23, 2019
Anna S. Roth, who said she used to work on commercial facial recognition technology, responded to Ocasio-Cortez's comments with an explanation of how algorithms can come to be biased:
1/ @AOC is completely on point. I used to work on face reco commercially (and having been on the wrong side of working on a biased model), I thought people might find helpful concrete examples of *some* of the ways face AI systems can come to be biased. https://t.co/s8Y5E7OBFC— Anna S. Roth (@AnnaSRoth) January 26, 2019
Emily Gorcenski, a data scientist, also backed Ocasio-Cortez up:
I’m a Lead Data Scientist for a major international consultancy. I have a degree in computational mathematics. I have 15 years experience. @AOC is right, this phenomenon is well-recognized in the field, and I’d enjoy hearing Ryan try to explain the Hahn-Banach theorem to me. https://t.co/BmgLLvUGBu— Emily G (@EmilyGorcenski) January 23, 2019
Other lawmakers have also engaged in tough talk on tech policy in recent years on a host of issues from biased facial recognition software to privacy. Sen. Ed Markey (D-Mass.), for instance, has been particularly active on facial recognition, and told me in December that we "currently lack key safeguards" to protect civil liberties from the risks associated with the technology.
But so far, the tough rhetoric on algorithmic bias hasn’t yet translated into laws on the books that would rein in the technology companies in a meaningful way -- and lawmakers seeking change could likely use all the help they can get.
Even lawmakers’ attempts to put technology executives in the hot seat in the wake of scandals have at times backfired, resulting in the lawmakers appearing to know little about technology. In one hearing with Google chief executive Sundar Pichai, lawmakers phrased their lines of questioning too generally to get real answers about important topics such as location data collection. In a hearing with Facebook chief executive Mark Zuckerberg, some lawmakers like former Sen. Orrin G. Hatch (R-Utah) seemed to not even understand how the companies make money.
Democrats controlling the House have promised to take on tech giants in the new Congress, and Ocasio-Cortez is clearly positioning herself to be part of that effort. This weekend, she also took a swipe at “tech monopolies," arguing that they’re killing journalism:
Gonna keep it 💯: The biggest threats to journalism right now are tech monopolies & concentration of ownership.— Alexandria Ocasio-Cortez (@AOC) January 26, 2019
Healthy democracy *requires* high-quality journalism.
W/o a wide range of independent outlets & the revenue to sustain them, our democracy will continue to crumble.
Correction: A previous version of this article referred to Orrin Hatch as a senator. He is a former senator.
|You are reading The Technology 202, our guide to the intersection of technology and politics.|
|Not a regular subscriber?|
BITS: The United States over the past year took on a campaign to limit the involvement of Chinese telecommunications giant Huawei and other Chinese companies in building 5G networks, the New York Times's David E. Sanger, Julian E. Barnes, Raymond Zhong and Marc Santora reported. Those next-generation networks are set to allow data to travel significantly faster and will help expand the use of artificial intelligence and virtual reality. As part of its campaign, the Trump administration has raised pressure on allies including Britain and Poland to ban Huawei from taking part in the development of their 5G networks.
“The administration contends that the world is engaged in a new arms race — one that involves technology, rather than conventional weaponry, but poses just as much danger to America’s national security,” according to the Times. However, some countries have wondered whether the Trump administration's efforts to isolate Huawei and other Chinese firms are indeed linked to national security concerns or if they seek to hinder China's efforts to compete economically with the United States.
NIBBLES: Two emergency departments in Los Angeles received more people who were injured while riding electric scooters than people who were injured while riding bicycles or traveling by foot between September 2017 and August 2018, according to a new study, The Washington Post's Peter Holley reported. The study, which was published in the medical journal JAMA Network Open, found 195 visits for bicyclist injuries to the two emergency departments, 181 visits for pedestrian injuries and 249 visits for injuries related to electric scooters. As electric scooter companies including Bird and Lime have expanded to numerous cities across the United States, some health professionals have called the resulting injuries a “public health crisis,” according to my colleague.
In the study, almost 28 percent of the people with injuries related to electric scooters had contusions, sprains and lacerations. “About 30 percent had fractures, and just over 40 percent were treated for head injuries, the study found,” Peter reported. “Nearly all the patients were discharged from the emergency departments, but 15 were admitted to a hospital, including two with severe head injuries who were placed in intensive care units.”
BYTES: Youtube said it is changing its recommendation algorithm to avoid promoting conspiracies and false information in a sign the company is trying to curb disinformation on the platform, The Post's Elizabeth Dwoskin reported. The change will be gradual and will apply to less than 1 percent of the content on the platform. It will also affect English-language videos only. The company said in a blog post that it is “taking a closer look at how we can reduce the spread of content that comes close to — but doesn’t quite cross the line of — violating our Community Guidelines.”
For instance, YouTube said it will reduce recommendations for videos that claim that the Earth is flat or that make “blatantly false claims about historic events” such as the Sept. 11, 2001, terrorist attacks. However, the videos would not be deleted from YouTube. “YouTube, which has historically given wide latitude to free speech concerns, does not prohibit conspiracy theories or other forms of false information,” Elizabeth wrote. “The company does ban hate speech but defines it somewhat narrowly as speech that promotes violence or hatred of vulnerable groups.”
— A plan by Facebook chief executive Mark Zuckerberg to integrate WhatsApp, Instagram and Facebook Messenger could raise questions about user privacy, the New York Times's Mike Isaac reported. The initiative also represents a change in Zuckerberg's stance after he had said that WhatsApp and Instagram would enjoy broad autonomy when Facebook acquired the companies. Under the plan, the apps would share a unified infrastructure but they would continue to operate separately, according to the Times. “By stitching the apps’ infrastructure together, Mr. Zuckerberg hopes to increase Facebook’s utility and keep users highly engaged inside the company’s ecosystem,” Isaac reported. “That could reduce people’s appetite for rival messaging services, like those offered by Apple and Google.”
Rep. Ro Khanna (D-Calif.) criticized the plan on Twitter:
2/2 -- Imagine how different the world would be if Facebook had to compete with Instagram and WhatsApp. That would have encouraged real competition that would have promoted privacy and benefited consumers.— Rep. Ro Khanna (@RepRoKhanna) January 25, 2019
— Some tech leaders at the World Economic Forum in Davos, Switzerland, this year said that “they have a greater responsibility to mitigate the dangers of their own creations” as major tech firms face growing criticism, the Wall Street Journal's Sam Schechner reported. “For years, tech executives met with business leaders and government officials here in Davos and other venues, projecting unbridled confidence that their technologies were making the world a more open, prosperous place,” according to the Journal. “Now, while maintaining their optimism that tech is a force for good, companies are embracing the idea of new regulations and talking about greater social obligations.”
— More technology news from the private sector:
— The European Commission said that European privacy regulators have received more than 95,000 complaints about possible data breaches since the General Data Protection Regulation went into effect, Reuters reported. Regulators have opened 225 investigations. “The majority of the complaints focused on telemarketing, promotional emails and video surveillance by closed-circuit televisions,” according to Reuters.
— The Illinois Supreme Court dismissed a case that would have curbed the power of a 2008 state law that limits the use of biometric markers, the Verge's Russell Brandom reported. “Whatever expenses a business might incur to meet the law’s requirements are likely to be insignificant compared to the substantial and irreversible harm that could result if biometric identifiers and information are not properly safeguarded,” according to the court's ruling. Facebook and Google have faced lawsuits over alleged violations of Illinois's Biometric Information Privacy Act in their photo-tagging product, according to the Verge.
— More technology news from the public sector:
— More news about tech workforce and culture:
— Tech news generating buzz around the Web:
— News about tech incidents and blunders:
- IoT Evolution Expo in Fort Lauderdale, Fla., tomorrow through Friday.
- Upfront Summit in Malibu, Calif., on Wednesday through Thursday.
- ITEXPO in Fort Lauderdale, Fla., on Wednesday through Friday.
Lawmakers have three weeks to make a deal on border security:
The colorful, quirky and weird parts of the Stone indictment, explained:
Man arrested after alligator discovered in his D.C. home: