bias
Latest
Utah pauses Banjo's AI surveillance after learning of owner's racist past
Utah authorities have put use of Banjo's surveillance tech on hold while they investigate for bias due to the owner's racist past.
Court finds algorithm bias studies don't violate US anti-hacking law
Researchers should be free to look for bias in website algorithms in the future. A federal court in DC has ruled in a lawsuit against Attorney General William Barr that studies aimed at detecting discrimination in online algorithms don't violate the Computer Fraud and Abuse Act. The government argued that the Act made it illegal to violate a site's terms of service through some investigative methods (such as submitting false info for research), but Judge John Bates determined that the terms only raised the possibility of civil liability, not criminal cases.
Researchers highlight racial bias in speech recognition systems
Researchers have identified significant racial disparities in speech recognition systems from five of the world's biggest tech companies. According to a study from Stanford University, systems from Amazon, Apple, Google, IBM and Microsoft make far more errors with users who are black than those who are white.
Amazon wants to question Trump over his role in the $10 billion DoD contract
It's no secret that Amazon was unhappy with the Pentagon's decision to award the $10 billion Joint Enterprise Defense Infrastructure (JEDI) contract to Microsoft. Now, Amazon wants to depose President Trump and six other officials, CNBC reports. The company hopes to question Trump about his involvement in the bidding process, including private conversations, instructions given related to the award and "efforts to harm Amazon and AWS."
Your online activity is now effectively a social ‘credit score’
Kaylen Ward's Twitter fundraiser for the Australian bushfire relief has ended. The Los Angeles-based model said she raised $1 million (by comparison Jeff Bezos donated $690,000). At the start of Ms. Ward's successful donation drive she had three Instagram accounts — none of which were part of the campaign.
Study says YouTube 'actively discourages' radicalism
Politicians and others complain that YouTube fosters extremism, but how caustic is it, really? Not all that much, according to researchers. Data scientist Mark Ledwich and UC Berkley researcher Anna Zaitsev have published a study suggesting that YouTube "actively discourages" radicalism through its recommendation system. Their reviewers classified over 760 politics-oriented channels based on overall leaning, topics and proximity to the mainstream, and found that YouTube removed "almost all" suggestions for conspiracy theorists, white identitarians and "provocateurs" (read: purposefully offensive creators). For the most part, there's only a significant likelihood of being matched with questionable content if you're already watching that material.
Federal study shows face recognition accuracy varies by gender and race
Researchers have studied the potential for bias in facial recognition algorithms before, but now it's the US government's turn to weigh in. The National Institute of Standards and Technology has published a study indicating "demographic differentials" in the majority of the facial recognition algorithms it tested. The report, which examined both one-to-one matching (such as verifying a passport photo) and one-to-many matching (looking for criminals in a crowd), saw noticeable surges in false positives based on gender, age and racial background -- but cautioned against this representing definitive proof of systemic bias.
Healthcare risk algorithm had 'significant racial bias'
There's more evidence of algorithms demonstrating racial bias. Researchers have determined that a "widely used" risk prediction algorithm from a major (but unnamed) healthcare provider had a "significant racial bias." While it didn't directly consider ethnicity, its emphasis on medical costs as bellwethers for health led to the code routinely underestimating the needs of black patients. A sicker black person would receive the same risk score as a healthier white person simply because of how much they could spend.
An independent report on Facebook’s alleged liberal bias tells us nothing
Conservatives have long lamented that Facebook has a liberal, or anti-conservative, bias. Since the 2016 election, the company has been grilled on the issue by the White House more than once. In an attempt to clear the air, Facebook enlisted an independent third-party to decide once and for all if it does indeed have an anti-conservative bias. Last year, it asked former Republican Senator Jon Kyl and his team at Covington & Burling LLP to conduct a review of the company's policies. The results are in, and for the most part, they tell us nothing.
Oakland bans city use of facial recognition software
Oakland, Calif. is now the third US city to ban the use of facial recognition technology. The city passed an ordinance yesterday that prohibits the use of the technology on the grounds that it is often inaccurate, potentially invasive and lacks standards. "Face recognition technology runs the risk of making Oakland residents less safe as the misidentification of individuals could lead to the misuse of force, false incarceration, and minority-based persecution," Council President Rebecca Kaplan wrote in a letter recommending the ban.
Axon won’t use facial recognition tech in its police body cameras
Axon, a major supplier of police body cameras and software, announced today that it will not include face-matching technology in its body cameras -- at least not yet. The decision follows a report from Axon's independent AI ethics board, which concluded that face recognition technology is not reliable enough to justify its use in body cameras. According to the report, there is "evidence of unequal and unreliable performance across races, ethnicities, genders and other identity groups."
UN study finds female voice assistants reinforce harmful stereotypes
For the record, the appropriate response to being called a slut isn't, "I'd blush if I could." But that's what Siri is programmed to say. According to a report by the United Nations, the fact that most voice assistant are gendered as young women is reinforcing harmful stereotypes that women are docile and eager to please, even when they're called lewd names.
White House requests reports of 'censorship' and your email address
Following up on political arguments over charges of "censorship" on social media, the White House has linked to a questionnaire on TypeForm.com where it's collecting reports of bias. In a tweet, the White House encouraged entries "no matter your views" although it doesn't guarantee any response to submissions.* (If you've been suspended for explosive statements like "Boyz II Men Are Trash", then you may also want to try contacting the EFF -- or just contact the EFF, whatever you prefer.)
Artificial intelligence is on the brink of a 'diversity disaster'
The lack of diversity within artificial intelligence is pushing the field to a dangerous "tipping point," according to new research from the AI Now Institute. It says that due to an overwhelming proportion of white males in the field, the technology is at risk of perpetuating historical biases and power imbalances.
Senate bill would make tech companies test algorithms for bias
It's well established that algorithms can exhibit bias, however inadvertently, and a trio of US politicians believe they can do something about it. Senators Cory Booker, Yvette Clarke and Ron Wyden have introduced an Algorithmic Accountability Act that would require larger companies to test their algorithms and fix anything "inaccurate, unfair, biased or discriminatory." The move would also ask them to study how their systems protect personal data,and would let the FTC create regulations mandating impact studies for "highly sensitive" automated systems.
Tech companies spent more than $64 million on lobbying in 2018
The biggest companies in tech spent more money lobbying the government in 2018 than in any year previous. According to a report from Reuters, Google dropped $21.2 million on lobbying efforts in 2018, a new record for the company and the most it has spent on lobbying since 2012. Likewise, Facebook cranked up its lobbying efforts by dropping $12.62 million, the most the social network has ever spent on political influence.
Microsoft details principles that guide its facial recognition work
In July, Microsoft President Brad Smith published a blog post discussing some of the challenges and risks surrounding facial recognition technology and how his company plans to address them. He said Microsoft was working on developing a set of guiding principles to direct the development of its own technology and this week, he shared what those principles are. In a new blog post, Smith described six principles that he says will govern Microsoft's work while it continues to advocate for legislation that will more concretely guide this technology as a whole.
Google is working to remove gender bias in its translations
Google has taken a step towards reducing gender bias on the Google Translate website. In the past, when a word could be translated in either a masculine or feminine form, only one translation was provided, and because Google Translate learns from existing examples of translations, biases in those samples can be transferred to Translate's responses. That could sometimes result in words like "strong" or "doctor" leading to masculine translations while words like "nurse" or "beautiful" could produce feminine translations. Now, however, for certain languages, Translate will offer both a masculine and a feminine translation when either might be appropriate.
Google CEO will now testify before Congress December 11th
Google CEO Sundar Pichai was scheduled to appear before the House Judiciary Committee Wednesday, but that's no longer the case following the death of George H. W. Bush. The former president's funeral will take place that day and, as such, Pichai's hearing has been rearranged for December 11th at 10 AM ET.
Google CEO Sundar Pichai will testify in Congress on December 5th
Republican politicians, most notably US President Donald Trump, have strongly criticized Google for bias against conservatives. "There's a lot of interest in their algorithm, how those algorithms work, how those algorithms are supervised," said outgoing committee member Bob Goodlatte in September. To counter that, CEO Sundar Pichai is scheduled to appear on December 5th before the House Judiciary Committee to dispel any idea of bias in its all-powerful search algorithms, according to the Washington Post.
