Facebook users who watched a newspaper video featuring black men were asked if they wanted to “keep seeing videos about primates” by one of the platform’s artificial intelligence recommendation systems. It’s only the latest in a long-running series of errors that have raised concerns over racial bias in AI.
Facebook told the BBC it “was clearly an unacceptable error,” disabled the system and then proceeded to open an investigation. “We apologise to anyone who may have seen these offensive recommendations,” continued the social media giant.
Back in 2015, Google came under fire after its new Photos app categorised photos in one of the most racist ways possible. On 28 June, computer programmer Jacky Alciné found that the feature kept tagging pictures of him and his girlfriend as “gorillas.” He tweeted at Google asking what kind of sample images the company had used that would allow such a terrible mistake to happen.
“Google Photos, y’all fucked up. My friend’s not a gorilla,” read his now-deleted tweet. Google’s chief social architect Yonatan Zunger responded quickly, apologising for the feature: “No, this is not how you determine someone’s target market. This is 100% Not OK.”
The company said it was “appalled and genuinely sorry,” though its fix, Wired reported in 2018, was simply to censor photo searches and tags for the word “gorilla.”
In July 2020, Facebook announced it would be forming new internal teams that would determine whether its algorithms are racially biased or not. The recent “primates” recommendation “was an algorithmic error on Facebook” and did not reflect the content of the video, a representative told BBC News.
“We disabled the entire topic-recommendation feature as soon as we realised this was happening so we could investigate the cause and prevent this from happening again. As we have said while we have made improvements to our AI, we know it’s not perfect and we have more progress to make,” they continued.
Research has previously shown that AI is often biased. AI systems aren’t perfect—over the last few years, society has begun to grapple with exactly how much human prejudices can find their way through AI systems. Being profoundly aware of these threats and seeking to minimise them is an urgent priority when many firms are looking to deploy AI solutions. Algorithmic bias can take varied forms such as gender bias, racial prejudice and age discrimination.
First and foremost, however, we need to admit and recognise that AI isn’t perfect. Developing more inclusive algorithms—with a specific goal of removing social bias—is the only way we can prevent this from happening again. Until then, AI will continue to make racist mistakes—driven by human error—for the years to come.
In today’s media landscape, people are bound to be trapped in virtually inescapable echo chambers. As social media is now our primary source of news and information, we find ourselves at the mercy of algorithms that are designed to feed us with content we’d find ‘agreeable’. This problem transcends social media, though, as each news source now has its own political agenda and coverage, be it Fox News or CNN.
Even those of us who are painfully aware of this issue are struggling to do something to resolve it. Fact-checking every single news item on our own is a laborious task that would pretty much count as a full-time job, and finding information sources that are both credible and challenging is hard seeing as our virtual surroundings become increasingly homogeneous. That’s where Ground News comes in.
Ground News, the world’s first news comparison platform, helps readers compare how the same news item was covered in various sources across the political spectrum and around the globe, and encourages them to reach our own conclusions. Ground News currently features over 50,000 news sources, all of which were vetted for credibility.
Take the issue of the ongoing Democratic primaries, for instance, or even Megxit—Ground News will map out the various sources that covered these stories and categorise them based on the publication’s political bias, which are labelled as ‘lean left’, ‘left’, ‘centre’, ‘lean right’ and ‘right’. Through such coverage analysis, a reader could easily get a balanced and diverse picture of each event and get access to points of view they would have not been exposed to otherwise. Readers can consult sources as left-leaning as Mother Jones and right-leaning as Breitbart, and get a glimpse into rare coverage from places like Iran and North Korea (all of which had been vetted using tech).
Harleen Kaur, Ground News’ CEO and co-founder, comes from an engineering background. She used to work for NASA, and later on at a satellite startup where her team used earth-observation satellite technology to help, among other things, insurance companies identify fraud and municipalities spot leaking pipes.
Following the 2016 elections, when fraudulent reports flooded the internet and hindered the democratic process, Kaur felt compelled to help rectifying our media ecosystem. “I [couldn’t] believe, as an engineer, that there’s so much tech around us and yet we don’t know what to believe anymore,” Kaur told Screen Shot. “I wanted to present a solution where somebody who was not trained as a journalist, somebody who doesn’t spend their day understanding what really happens in the news, will have a simple tool to be able to do that, and read news from various sources.”
Once a week, Ground News publishes a Blindspot Report, which lists the news that were excluded from certain news sources. A left-leaning American, for instance, could have easily missed the news about Mike Bloomberg considering Hillary Clinton as his potential Vice President, whereas readers leaning right were most likely oblivious to the fact that the Chinese authorities were aware of the coronavirus threat before reports about it became public. “Whenever there are climate catastrophes, the right-wing media is mostly quiet on it, or if something adverse happens with the immigration situation—the right-wing media is quiet on it,” Kaur said, “The left does the same as well on certain topics. And we want to show the consumers that it doesn’t mean that if their news source isn’t reporting something, then it’s necessarily not true.”
Other advantages unique to Ground News are the time and location features, which allow readers to see how the coverage of a certain topic changes over time and depending on the geographic location of the publication. Reading about the downing of the Ukrainian passenger aircraft by the Iranian authorities, for instance, readers were presented with a timeline of the events as they unfolded, and how they were covered differently by Canadian and Iranian news.
Currently, the majority of Ground News’ readers are located in the US, and Kaur says that their intention is to focus primarily on the American market for now, considering it is a tumultuous election year. That said, the news comparison platform is seeing a steadily growing readership in the UK—a country which Kaur believes “is another place where biases are a bit more obvious in terms of media landscape,” as well as several other countries across the world.
There is no quick fix to our contaminated media environment, and it is up to each of us to broaden our lenses and diversify our intake of information and views. A platform like Ground News can certainly support us on this quest. Are you ready to burst your bubble?