Humans

Computers are worse at recognising women and people of colour than white men. Documentary Coded Bias shows that the problems don’t stop there

Technology 12 August 2020
By Vijaysree Venkatraman
Face-recognition AI could only “see” Joy Buolamwini when she wore a white mask
7th Empire Media

Coded Bias
Shalini Kantayya
Ongoing film festival screenings

Advertisement

IN HER first semester as a graduate student at the MIT Media Lab, Joy Buolamwini encountered a peculiar problem. Commercial face-recognition software, which detected her light-skinned classmates just fine, couldn’t “see” her face. Until, that is, she donned a white plastic mask in frustration.
Coded Bias is a timely, thought-provoking documentary from director Shalini Kantayya. It follows Buolamwini’s journey to uncover racial and sexist bias in face-recognition software and other artificial intelligence systems. Such technology is increasingly used to make important decisions, but many of the algorithms are a black box.
“I hope this will be a kind of Inconvenient Truth of algorithmic justice, a film that explains the science and ethics around an issue of critical importance to the future of humanity,” Kantayya told New Scientist.
The documentary, which premiered at the Sundance Film Festival earlier this year, sees a band of articulate scientists, scholars and authors do most of the talking. This cast primarily consists of women of colour, which is fitting because studies, including those by Buolamwini, reveal that face-recognition systems have much lower accuracy rates when identifying female and darker-skinned faces compared with white, male faces.
Recently, there has been a backlash against face recognition. IBM, Amazon and Microsoft have all halted or restricted sales of their technology. US cities, notably Boston and San Francisco, have banned government use of face recognition, recognising problems of racial bias.

People seem to have different experiences with the technology. The documentary shows a bemused pedestrian in London being fined for partially covering his face while passing a police surveillance van. On the streets of Hangzhou, China, we meet a skateboarder who says she appreciates face recognition’s convenience as it is used to grant her entry to train stations and her residential complex.
“If an AI suspects you are a gambler, you could be presented with ads for discount fares to Las Vegas”
The film also explores how decision-making algorithms can be susceptible to bias. In 2014, for example, Amazon developed an experimental tool for screening job applications for technology roles. The tool, which wasn’t designed to be sexist, discounted résumés that mentioned women’s colleges or groups, picking up on the gender imbalance in résumés submitted to the company. The tool was never used to evaluate actual job candidates.
AI systems can also build up a picture of people as they browse the internet, as the documentary investigates. They can suss out things we don’t disclose, says Zeynep Tufekci at the University of North Carolina at Chapel Hill in the film. Individuals can then be targeted by online advertisers. For instance, if an AI system suspects you are a compulsive gambler, you could be presented with discount fares to Las Vegas, she says.
In the European Union, the General Data Protection Regulation goes some way to giving people better control over their personal data, but there is no equivalent regulation in the US.
“Data protection is the unfinished work of the civil rights movement,” said Kantayya. The film argues that society should hold the makers of AI software accountable. It advocates a regulatory body to protect the public from its harms and biases.
At the end of the film, Buolamwini testifies in front of the US Congress to press the case for regulation. She wants people to support equity, transparency and accountability in the use of AI that governs our lives. She has now founded a group called the Algorithmic Justice League, which tries to highlight these issues.
Kantayya said she was inspired to make Coded Bias by Buolamwini and other brilliant and badass mathematicians and scientists. It is an eye-opening account of the dangers of invasive surveillance and bias in AI.
More on these topics: More

J
Netflix/
Watch
John Was Trying to Contact Aliens tells the story of John Shepherd, who spent 30 years trying to contact extraterrestrials by broadcasting music millions of kilometres into space. On Netflix from 20 August.
Read
The science transforming the way we learn sees Sanjay Sarma, head of open learning at MIT, join fellow researcher Luke Yoquinto to explain how scientific findings in wildly different fields are transforming the way we learn and teach.
Visit
Driverless: who is in control? is an excellent exhibition about autonomous vehicles at London’s Science Museum. The museum reopens on 19 August, and is extending this show until January 2021.
More on these topics: More

Playing Superliminal confuses your perception of space, but the trick wears thin quite quickly, says Jacob Aron

Humans 12 August 2020
By Jacob Aron
Superliminal has a creative relationship with perspective
Pillow Castle Games

Superliminal
Pillow Castle
Multiple consoles

Advertisement

I HAVE been having strange dreams recently. This may be due to the ongoing coronavirus pandemic – a survey in March found that people in the UK have been getting more sleep due to lockdown measures, and more sleep can have an effect on your dreams.
Or it might be that I have been playing Superliminal, a first-person game designed to mess with your head and your perception of space. It takes place entirely in dreams, with the unnamed character you play as participating in an experimental form of therapy called Somnasculpt administered by a Dr Glenn Pierce.
The story here is pretty light. As you pass through the game, you hear messages from Pierce and the AI that is running the dream therapy, with both getting increasingly agitated as you become lost in the dreamscape, but that is about it. The plot is essentially a set-up for very clever forced perspective and other optical illusions.
This is demonstrated early in the game, when you pick up a chess piece from a table. Place it down again and it has changed in size to match your perspective. If that sounds confusing, think about the classic tourist photo of people pretending to support the leaning tower of Pisa, and imagine you could actually shrink it down to hold it up for real. You can repeat the trick over and over, making objects tiny or gigantic.
You use this ability to pass through a series of surreal puzzle rooms, placing objects on pressure plates or making a wedge of cheese large enough to use as a ramp to a high door. Later levels add complications, such as needing to stand in a specific spot to transform an image stretched across a wall into an object you can pick up.

Developer Pillow Castle loves to mess with you, changing the “rules” of the game just as you have figured out how something works. But, ultimately, the forced perspective wears thin. Many of the game’s puzzles can be solved by picking up an object, holding it in the air and watching a larger version fall to the ground with a thud. This is fun the first few times – I actually ran out of the way, worrying that I was about to be squished by a giant chess piece – but it doesn’t offer enough variety.
“I actually ran out of the way, worrying that I was about to be squished by a giant chess piece”
The game is obviously inspired by Portal, a 2007 release that kicked off the first-person puzzle genre, in which you also navigate a series of rooms while listening to an AI, in that case, the malevolent GLaDOS, which berates you at every turn. Rather than forced perspective, you use a “portal gun” to solve puzzles. This allows you to connect two surfaces via a wormhole through which you and objects can pass.
In later levels, Superliminal introduces its own version of portals in the form of linked doorways that can be resized, making you grow or shrink as you pass through them. It is a fun idea, but in practice I found it very fiddly. My struggles to line up the doors in the way I wanted left me pining for Portal’s elegance.
It is perhaps unfair to compare Superliminal to one of the greatest games of all time, but it doesn’t help itself by aping Portal so closely. The game does at least have a more optimistic tone than Portal‘s cynicism, ending with a positive message that some people may find to be a genuinely useful takeaway from the experience.
More on these topics: More

Greeting neighbours or gossiping with a colleague can boost your health and well-being, but coronavirus lockdowns are putting that in jeopardy. Here’s how to stay connected

Health 12 August 2020
By David Robson
AT THE beginning of the UK lockdown, I woke each morning with a feeling of impending doom. I was scared about covid-19, of course, but also worried about isolation. How would I cope without seeing friends and family? How could I perform my job as a journalist if I couldn’t meet people?
These weren’t baseless fears. In recent decades, a raft of research has shown that individuals with richer social worlds tend to have better mental well-being and lower stress, and to perform better at work. Missing out on our interactions with friends, colleagues and even shopkeepers can have a surprisingly powerful impact on our health.
WhatsApp conversations and Zoom “parties” have helped me to maintain a sense of connection, but these tools can’t replace aspects of interaction – like social touches and impromptu chats by the water cooler – that can boost mood and strengthen relationships.
Microsoft CEO Satya Nadella suggested as much in a recent interview with The New York Times. Although he felt the shift to digital interactions was going relatively smoothly, he wondered if we were burning through the “social capital” built up over years. He suspected that social bonds might start to evaporate. “What I miss is when you walk into a physical meeting, you are talking to the person that is next to you, you’re able to connect with them for the two minutes before and after,” he said.
“A wealth of studies have shown that high ‘social capital’ enhances our quality of life”
As many of us continue to work remotely, the long-term effects of social distancing could be serious. What can … More

The nuclear bomb told us we are the greatest threat to our own survival – and the covid-19 pandemic shows the lessons still to learn, say Anders Sandberg and Thomas Moynihan

Health | Comment 5 August 2020
By Anders Sandberg and Thomas Moynihan
Nagasaki was hit by an atomic bomb three days after Hiroshima
Gamma-Keystone via Getty Images

ON 6 AUGUST 1945, a nuclear bomb was dropped on the Japanese port city of Hiroshima. Three days later, Nagasaki suffered the same fate. Three-quarters of a century on, the full human toll is still unclear. In Hiroshima alone, some 75,000 souls were obliterated instantly, with many more deaths in the following months and years.
These are the only times nuclear weapons have been used in war; debates about the rights and wrongs continue. As we remember those who died, we might also usefully … More

New research reveals that miscarriage serves a critical role in human evolution – and in some instances, may even be associated with optimal fertility

Health 5 August 2020
By Alice Klein
WHEN I saw the positive result on my at-home pregnancy test, my mind raced ahead. I imagined how it would feel to hold my child for the first time, what we would call them. I thought of the bedtime stories we would read, pictured family camping holidays at the beach.
I never imagined that, just weeks later, while dancing at a friend’s wedding, a sharp twisting pain would signal that the pregnancy was over.
Like many women who have a miscarriage, I worried I had done something to trigger the loss. Had I exercised too hard? Slept too little? Around the world, studies show that many women experience shame and guilt after losing a pregnancy. One US survey found that 40 per cent of women who had a miscarriage believed it was because of something they did wrong. Though there is no evidence covid-19 increases miscarriage risk, the pandemic only exacerbates these worries. Society can add to the problem. In some countries, the culture of blame is so widespread that losing a pregnancy can land a woman in jail.
When I looked into the latest research, what I discovered not only challenged ideas that women are somehow responsible for their miscarriages, or experience them because something is wrong, but suggested that, surprisingly, they are usually associated with optimal maternal health. With advances in fertility medicine, we are finally starting to understand what happens in a miscarriage. This progress may offer solace when pregnancies don’t work out and help women struggling to become pregnant. It could even shed light on the role of miscarriage in our evolution. … More

Feedback is our weekly column of bizarre stories, implausible advertising claims, confusing instructions and more

Humans 5 August 2020

Josie Ford

Cursed crescent
One time, on a particularly cloudy holiday with friends, Feedback glanced up at the sky and noticed a dim ball of light floating not far above the horizon.
“Look at that,” we said, for want of anything more interesting to say. “It’s the day-moon.”
Readers, it was not the day-moon. It was, in fact, the sun.

Advertisement

What this anecdote serves to illustrate is that if something has been askew in the heavens of late, Feedback would be among the last to notice.
So you can imagine our surprise at discovering that a coven of “baby witches” has hexed the moon. Or, at least, so says the internet.
It appears that in an occult corner of the social networking app TikTok – known as WitchTok to its friends – a group of young witches decided to cast a curse on the moon. This appears to have caused all sorts of turmoil within the witchcraft community, and no little amusement outside it.
For, after all, we people of science know that the moon cannot be hexed. The moon isn’t some primordial reservoir of arcane energy to be used in witchcraft. It is a symbol for mutually antagonistic countries to race towards in an attempt to prove the relative superiority of their way of life. Much more sensible.
Lean times
A worrying trend in the Feedback inbox of late is the amount of attention that nominative determinism spotters are devoting to New Scientist itself.

We pass no judgement on this, but point ominously at a drawing of a snake eating its own tail while muttering about infinite recursion under our breath.
This week, for example, James Haigh writes in to comment on the name of an expert quoted in an article on public health policy regarding obesity.
“Michael Lean interviewed for the ‘Public health’s hard problem’ article??” asks James, making excellent use of the lesser-spotted (well, double-spotted, really) double question mark. “You couldn’t make this stuff up.”
Rumbling on
Some weeks ago, Feedback invited readers to send in the opening lines of limericks that we would do our humble best to complete.
Thank you to Ted Webber for throwing down the first gauntlet, based – in his words – on a New Scientist cover story. The opening line he wanted us to riff off was “If consciousness lies in our gut”. Well, Ted, here you go. Don’t say you didn’t ask for it.
If consciousness lies in our gut,
Then what is the role of the butt?
Neither Kant nor Foucault
Have pretended to know,
But to us it seems: open and shut.
Don’t be a square
Big news for geometry fans this week, as a German court has ruled that the Ritter Sport brand of chocolate can keep its trademark on square-shaped bars.
In its report, the BBC referred to the case as reinforcing Ritter’s “three-dimensional monopoly”, which – while being a charming phrase – perplexed Feedback. It goes without saying that the chocolates are three-dimensional: to our knowledge, no one has yet derived any pleasure from licking an atom-thick layer of chocolate spread off a graphene substrate.
But the trademark specifically covers square chocolate, not cubic chocolate. This, we are afraid, is a two-dimensional monopoly. And the reason we are afraid to say it is because the last time we checked, Hasbro had the trademark on that.
New chip on the block
While we are on the subject of chocology (chocolatey topology), Feedback was intrigued by a story this week about the quest to redesign the chocolate chip.
It turns out that the conventional tear-drop shaped chocolate chip, while effective in a brute force sort of way, lacks the geometrical finesse that chocolate chip cookie bakers wish it would have.
Namely, according to The Times, “it lacks a broad surface area to maximise taste and texture”. That is why Remy Labesque at Tesla – yes, electric car maker Tesla – has spent three years attempting a chocolatey redesign.
The new shape is a squashed diamond that tapers in three directions to maximise the textures it can achieve when melted. It is aesthetic, allegedly scientific and above all tasty. Feedback will be awaiting future updates with heavy and bated breath.
Good knights
There have been times of late, what with all this plague business going around, that the world has seemed to take on a distinctly medieval hue.
If you find this state of affairs discomfiting, then Feedback’s suggestion is that you stay well away from the Swedish island of Gotland. According to a report in The Times, the powers that be on Gotland have commissioned a troupe of knights on horseback to patrol the area around the ferry terminal, reminding people to socially distance.
The article is sadly lacking in detail about how exactly these reminders are to be enforced. At the point of a lance, perhaps? Or through several layers of PPE chainmail? Either way, the convergence of Sweden, medieval knights and global pandemic has a certain The Seventh Seal-iness about it that is making Feedback shiver.
Got a story for Feedback?
You can send stories to Feedback by email at feedback@newscientist.com. Please include your home address. This week’s and past Feedbacks can be seen on our website. More

Solving the Conway knot problem took mathematician Lisa Piccirillo on a journey into the fourth dimension. Here’s how she did it

Humans 5 August 2020
By Chelsea Whyte

Rocio Montoya

OVER the course of one week in 2018, Lisa Piccirillo cracked a mathematical problem that had gone unsolved for half a century. Posed by legendary mathematician John Conway in 1970, it concerns a complex geometrical object known as the Conway knot. While an ordinary overhand knot – the kind you would tie at the end of a thread – sees the string cross over itself three times, the Conway knot has 11 crossings. What Conway wanted to know is whether his knot can be formed by cutting a slice out of a more complex four-dimensional knot – or, as mathematicians put it, is it “slice”?
Piccirillo discovered that it isn’t. Her breakthrough came after finding a back door into the problem that could help mathematicians understand other four-dimensional objects. Currently a post-doctoral mathematician at Brandeis University in Waltham, Massachusetts, solving the Conway knot – along with her other research – has seen her offered a tenure-track position at the Massachusetts Institute of Technology. New Scientist spoke to her about the week she spent on the problem, her approach to mathematics and why it is time we stopped talking about geniuses.
Chelsea Whyte: How did you first become interested in mathematics?
Lisa Piccirillo: As a kid, I always liked maths and I was good at it in school. I’m from quite a rural area in Maine, and people said “if you like maths, you can become an engineer”. So I thought that’s what you do with maths, become an engineer. I went to a lot of day camps for engineering and made a lot of bridges out of popsicle sticks, and found out that … More