Against a field of deep blue, sword‑bearing demons surge toward a sacred fire. Ram — yellow‑clad, rendered in blue — lets loose a rain of arrows, with Lakshmana steady at his side. Vermilion blood spills from the demons as the ritual tilts toward battle.

Demons Approaching Rama and Lakshmana at a Fire Ceremony (Homa), illustration from a Ramayana series, India, Rajasthan, 18th-19th century. Photo: © President and Fellows of Harvard College, courtesy Harvard Art Museums.

This 18th-19th century Rajasthani miniature, Demons Approaching Rama and Lakshmana at a Fire Ceremony, does not just capture a charged moment from the Ramayana where ritual and battle converge. It is a snapshot of global trade and cultural exchange in the 1800s.The dominant pigment in the painting is Prussian blue, synthesized in Berlin for the first time around 1706, manufactured across Europe and exported to Asia. In Japan, Hokusai used it to create The Great Wave (1831) – the iconic image that is a popular screensaver on digital devices today. The Indian yellow is a local pigment, once produced primarily in Bihar from the urine of cows that were fed only mango leaves. Before the British banned the production of Indian yellow, the pigment had made its way to Europe. In this Rajasthani painting, the Prussian blue and Indian yellow come together to create the dark green of the holy man’s sash and one of the demon’s shorts.

The chemical composition of the pigments used in this work and more than 200 other South Asian and Himalayan paintings and manuscript folios is now available at the click of a mouse, thanks to Harvard University’s Mapping Color in History initiative. Focusing on the materials and histories of South Asian art, its searchable database allows users — anyone, anywhere in the world with reliable internet access — to explore the paintings by title, pigment, keyword, date, region and more. The initiative, which began in 2018, is the product of a collaboration between conservators, computing specialists and art historians among others. The database offers a new way to read history – through color.

“The Great Wave”. Credit: Katsushika Hokusai, Public domain, via Wikimedia Commons.

It all began with a pigment puzzle. In 2016, a conservation scientist at a museum in Boston detected cobalt in a 15th-century Jain manuscript. But smalt, a cobalt-infused blue glass pigment, was manufactured in Europe and exported to Asia only from the 17th century. The assumption was that conservationists had retouched the color. It piqued the interest of Jinah Kim, professor of Indian and South Asian Art at Harvard University, who was gathering data on pigments as research for her book on the material history of Indian painting. Kim wondered, “Did everything have to come from Europe? What do we know about actual pigment usage in India at this time period?”

Katherine Eremin, Senior Conservation Scientist at Harvard Art Museums, cracked the cobalt pigment puzzle. Eremin narrowed down the origin of the glassy cobalt pigment using its geochemical signature – it had trace elements associated with a cobalt mine in Kashan in Iran. The deep-blue smalt pigment, assumed to be European, most likely originated in Asia. She also analyzed a 17th century Rajasthani painting in the Harvard collection and determined that it had smalt which originated in Europe.

A Nayika and Her Lover, page from a dispersed Rasamanjari (Blossom Cluster of Delight) series, India, Basohli, c. 1660-70. Credit: Harvard Art Museums/Arthur M Sackler Museum.

Kim’s larger question remained: was there an indigenous system of colorants in India we know nothing about?

Long before synthetic hues, artists sourced their colours from a range of mineral pigments, plant-based dyes and insect-derived colorants. To identify and analyze these pigments is to understand the material core of art. Consider the color green in Indian miniatures. Is it a blend of blue and yellow pigments, as in Demons Approaching Rama and Lakshmana at a Fire Ceremony? Or is the green from beetle wings? If copper is detected in a painting, does this vibrant green come from malachite or atacamite? A suite of microscopy, imaging and spectroscopic techniques, which use light to decode the molecular composition of materials, allows researchers to determine the chemical makeup of pigments. Such detailed analyses can guide conservation strategies, inform restoration choices and help authenticate valuable paintings.

Past and present

What you cannot learn from a dead artist, or glean through an analysis of paintings using instruments, you can ask a living practitioner of traditional art, says Kim. Today’s artisans often rely on recipes and techniques passed down through centuries. There is value in understanding the process. This concept of artistic traditions being passed on from generation to generation is enshrined in The Intelligence of Tradition in Rajput Court Painting, by Molly Emma Aitken.

In 2021, conservator Anjali Jain, the India-based research manager of the Mapping Color in History team, visited the workshop of Babulal Marotia, an award-winning miniature painter based in Jaipur, to document this living tradition. Marotia says he chooses pigments based on color and tone. Mixing colorants with gum Arabic gives them a smooth, glossy finish that allows for fine layering. Marotia paints on handmade paper with an ultra-fine brush, no longer made of squirrel hair.

As a postdoctoral fellow at Harvard’s Straus Center, material scientist Celia Chari analyzed forty‑two colorants from Marotia’s palette. A little over half aligned with pigments found in 16th‑century manuscripts: yellow harital (arsenic sulfide), red hinglu (mercury sulfide), and kajal (lamp black). Others had slipped out of use: the gentle red of safflower; the translucent maroon of lac, once derived from Kerria lacca insects and traded in Europe as Indian Lake. Ultramarine — lajwarda, the mineral blue once prized for its depth — has largely vanished, replaced by plant‑based indigo. “Limited availability and changes in modern supply chains make it difficult, even financially unviable, to source the exact materials used historically,” Jain says.

Among contemporary pigments, Chari found Prussian blue — a Rajasthani staple for more than two centuries — alongside synthetic stand‑ins for Indian yellow, such as chrome yellow. She also identified white arsenolite, an inorganic compound undocumented in South Asian art, and brown barium ferrite, with no recorded use as a pigment in any culture. If the presence of arsenic or mercury raises concern, Chari notes that Marotia is fully aware of the toxicity and relies on thorough handwashing as protection. In the end, it is aesthetics — not chemistry — that guides his choices.

Through this pioneering study, which was published earlier this year, the researchers have just begun to explore the materials of contemporary South Asian artists through a scientific lens. Such studies can help future conservation efforts of contemporary paintings and deepen our understanding of traditional pigments. “It is all part of a continuum,” said Kim.

Better conservation

Because India holds most of the surviving paintings– in its various styles– these institutions become natural partners in tracing how pigments travelled through centuries and regions. Most museums in India don’t have in-house instrumentation – a notable exception is the Chhatrapati Shivaji Maharaj Vastu Sangrahalaya, Mumbai.

For most museums, transporting artwork to instrumentation centres is fraught with risk. To make the analytical techniques accessible on site, the Mapping Color in History established the Mobile Heritage Lab equipped with portable, handheld instruments. Already, the mobile lab has been deployed for the pigment analysis of old illustrated manuscripts at the Asiatic Society in Mumbai. Similar studies are underway at the Maharaja Sawai Man Singh II Museum in Jaipur’s City Palace. Discussions are in progress for a similar collaboration with the Government Museum and Art Gallery in Chandigarh, home to a collection of antique Pahari paintings.

“Sometimes old objects of art are treated with such absolute reverence – they’re left alone when, in fact, research-based conservation efforts would serve them better,” said Kim. As Project Director of Mapping Color in History, she emphasises the need for analytical studies to guide informed decisions about conservation. While conservation itself lies outside the scope of the Mapping Color in History project, the goal is to offer science-based insights to museum conservators – empowering them to make evidence-based choices that enhance the longevity of the works in their care.

The more researchers probe these pigments, the clearer it becomes that every brushstroke carries a record of the world that made it. The surprise is not that the colors endure — it’s how much they’ve been waiting to tell us.

A Latin American equivalent of the same concept; tracing the materials and art in the 16th-18th century paintings in the Viceroyalties

Materiality between Art, Science, and Culture in the Viceroyalities (16th–18th Centuries) is a series of research seminars organized by the Universidad Nacional de San Martín in Argentina that is bringing together scholars to examine the artistic materials—everything from wood to feathers to animals and insects ground up to make pigment—used during this period in Latin America, what made them unique, and what they can tell us about the history of art and cultural exchange.

Goal: To rewrite colonial art history through the lens of materiality—foregrounding indigenous knowledge, ecological entanglements, and transatlantic exchanges.

The post Vibrant Pigments, Global Exchange appeared first on Vijee Venkatraman.

In 1883, before a packed house in Bengal’s Serampore College, with an audience that included the American Consul General, Anandibai Joshi, 18, declared her intention: “I go to America because I wish to study medicine,” she said, speaking in English before the College Hall. “Ladies both European and Native are naturally averse to expose themselves in cases of emergency to treatment by doctors of the other sex. In my humble opinion there is a growing need for Hindu lady doctors in India, and I volunteer to qualify myself for one.” Her decision came at a high cost. The townsfolk disapproved of an upper-caste Brahmin woman crossing the forbidden “black waters” and created scenes at the post office where her husband Gopal Rao Joshi worked. Progressive beyond his time, Gopal had encouraged his wife’s ambition to become a physician when their infant son died years earlier for lack of medical attention.

The post First Desi Woman Graduate Student in the US appeared first on Vijee Venkatraman.

Illustration by Islenia Mil for Science

Poacher

A gunshot pierces the skull of a tusker, an adult male elephant prized for its ivory, and it slowly slumps to the ground. This gruesome yet gripping opening shot launches the web series Poacher, inspired by a true story, in which a motley team of wildlife crime fighters exposed the largest ivory ring in Indian history.

The series is set in 2015 in the jungles of Kerala, an Indian state nicknamed “God’s Own Country,” where, in the previous decade, authorities had quashed an ivory smuggling ring involving transnational crime syndicates. Since then, surveillance technology has become commonplace and large‑scale poaching is thought to be a thing of the past. So, in the opening scene, when a whistleblower comes forward to offer information on “Raaz,” a dangerous elephant poacher, state officials are initially dismissive.

In this fast‑paced investigative procedural—the first three episodes of which debuted at Sundance—the narrative momentum is maintained without forfeiting character depth. India’s religious diversity and its multilingualism are on casual display: viewers will hear Malayalam, English, and Hindi spoken by different characters.

The show’s protagonists include computer programmer and snake expert Alan Joseph (played by Roshan Mathew) who builds a case against the poacher that “will live and die on data analysis,” and forest officer Mala Jogi (Nimisha Sajayan), who leads raids into the hideouts of dangerous suspects and cleans up after the botched efforts of her colleagues. When the team eventually tracks down Raaz, viewers realize this is just the beginning of a very complicated case.

Cleverly interspersed shots from the misty jungles suggest that the region’s animals too are keeping a wary eye on the proceedings. If the elephants go, the jungle ecosystem will collapse, and Kerala will eventually be as polluted as the national capital New Delhi, viewers are told. An aerial shot of the vehicle‑clogged arteries of that megacity hints at what would be lost if this came to be.

Wildlife crime fighters are overworked, with little personal time. In Poacher, their triumphs and struggles are told with empathy. The dedication of these men and women to this dangerous work suggests there is still hope for the future of wildlife on a planet where humans are now the top predators.

The Longest Goodbye

In the next decade, NASA plans to send astronauts to Mars on a three-year mission. The journey itself will take approximately six months, each way. While the various components of a spaceship can be tested under extreme conditions, the effect of prolonged social isolation on the crew members’ emotional well-being remains unknown. And yet, how well the astronauts hold up mentally and emotionally within those cramped quarters could make or mar the mission. This simple but profound idea is elegantly explored in The Longest Goodbye.

The documentary features interviews with Dr. Al Holland, a NASA psychologist who is tasked with keeping space explorers mentally fit throughout their missions, as well as insightful interviews with astronaut Cady Coleman who lived aboard the International Space Station (ISS) for 6 months from 2010 to 2011, Sukjin Han, a member of an Earthbound Mars simulation crew, and Kayla Barron, an astronaut currently in training for a potential Mars mission.

Archival video of Coleman’s interactions with her family—which include a long-distance musical duet and a game of tic-tac-toe— over a shaky internet connection during her six-month stint at the ISS, makes for heartwarming scenes.  Coleman’s son Jamey, then in 4^th grade, had a tough time with his mother’s absence though. As he explains in the film, he always tried to put on a brave face for her.

“Crew members’ connection with family is a critical piece of sustenance for them,” Holland observes. Such connections are important during any long period of separation, learned Holland in 2010, when NASA was called in to help manage the mental health of 33 Chilean miners trapped underground. The documentary includes footage of the miners’ 69 day ordeal and celebrated rescue, along with touching scenes of the miners’ families interacting with them through video calls.

In a Mars expedition, astronauts will not be able to communicate with their families in real-time, so experts are trying to come up with new strategies to counter homesickness. In the film, they discuss possible solutions, including virtual reality rendezvous with loved ones, AI-enabled companions, and even the possibility of inducing hibernation during the flight. A medical coma may spare the astronauts some angst en route, but will likely lead to readjustment issues when they awaken, they concede.

Sometimes, the mission to Mars feels like too much to ask of any human for the sake of science. And yet, the explorers who volunteer for such endeavors are often among the most eager participants. “If I could have spent another six months [on the ISS], I would have stayed in a minute,” reveals Coleman in the closing moments of the film.

Read here. SUNDANCE 2023

The post Science at Sundance 2023 appeared first on Vijee Venkatraman.

R.I.P Jim Watson. 

Twenty years or so ago, I worked on the Human Genome Project. In the Human Genome Project, a “finisher” was responsible for ensuring the accuracy, completeness, and continuity of DNA sequences—essentially polishing the raw genomic data into a reliable reference for researchers. All day, I scrutinized eye-glazing stretches of A’s, G’s, C’s and T’s, cloned bits from a composite human genome, corrected base-calling mistakes, misassemblies and other artifacts introduced during automated sequencing — closed gaps and deposited the polished versions of those genomic bits in GenBank every evening.  Anywhere in the world, a scientist could then begin to interpret the biology of the sequence. It was work which would be built upon in the decades to come.

This year, a magazine asked me to do a sum up of what the completion of the Human Genome Project meant. 

On April 25, 1953, the journal Nature published a one-page paper by two scientists from Cambridge University in the UK, James Watson and Francis Crick. The paper, titled ‘A Structure for Deoxyribose Nucleic Acid,” went on to set things in motion for a revolution in biology.  The structure of DNA, the authors wrote, suggests “a possible copying mechanism for the genetic material.”

While the chemical composition of DNA was well-known, its structure was not. Watson and Crick had shown that the Deoxyribose Nucleic Acid (DNA) was a double helix. DNA is a polymer made of two strands of molecules called nucleotides. The sugar and phosphate parts of the nucleotide form the two strands of the helix, and the nucleotide bases point into the helix, where they stack on top of each other. DNA molecules have four kinds of nucleotide bases. These bases pair with great specificity. Adenine (A) pairs with Thymine (T). Guanine (G) pairs with Cytosine (C). This pairing is key – it is the basis by which DNA molecules are copied when cells divide. In humans, DNA is packaged into 23 pairs of chromosomes – one from each parent. Each chromosome has its share of genes – the functional units of heredity. The genome is the sum of all the DNA in the nucleus.

Crick was the first to realize that the seemingly random sequence of the four bases in the genomic DNA formed a code and provided a template for protein synthesis. Other scientists would finally “break the genetic code” and describe how three nucleotide bases in a DNA code for each of the twenty amino acids, which are the fundamental building blocks from which all life is constructed. Crick, however, did not foresee that entire genomes would be decoded.

The DNA sequence provide the blueprint for development from a single cell to a complex, integrated organism. Determining the entire genomic sequence could help scientists gain molecular-level insights into the workings of any organism, but, at that point, sequencing entire genomes was unthinkable. It took a series of advances in molecular biology, technology, and computing to first make the sequencing of whole genomes, large and small, a reality.  

The DNA sequence of a virus, with less than 10,000 base pairs, would be published a full fifteen years after the discovery of the double helix. In 1977, Frederick Sanger devised a method to sequence DNA. Sanger and his colleagues determined the genetic sequence of the Bacteriophage phiX174, which had 5368 base pairs.  A rough estimate indicated that, without automation, it would take 1,500 scientists working for a century to sequence the human genome which is some 3 billion base pairs long. But the Sanger method was automated, and the first commercial sequencing machines hit the market in 1986.

By the mid-1980s, some visionary scientists proposed the idea of sequencing the entire human genome. Many agreed, in principle, that it would be useful to determine the order and spacing of all the genes that make up the genome, but some biologists thought this was too ambitious a project, and that it would end up generating plenty of useless data. Still, there was no denying the fact that the large-scale discovery of disease-causing genes would help in medical research and clinical care. The grant-making agencies greenlit the big science project. 

The Human Genome Project (HGP) officially began in 1990. Watson became the first leader of the U.S. government’s effort to sequence the human genome, the three billion base pairs contained in a cell. The National Institutes of Health (NIH) in the United States estimated that the project would take 15 years to complete at a cost of $3 billion. Select labs across the globe joined in and formed an international consortium. Apart from sequencing the human genome, their goal was also to identify all the genes it contained – the estimated number was 100,000 genes.

In the first phase of the Human Genome Project, scientists began sequencing model organisms, which had long been used in the lab, such as the bacterium Escherichia coli and the yeast Saccharomyces cerevisiae. The consortium published the sequences of the bacteria and the yeast in 1996. A year before that, Craig Venter, a maverick scientist, came along, completed the sequencing of the bacterium Haemophilus influenza — it was first free-living organism whose genome was sequenced. Meanwhile, the consortium lay the groundwork to sequence the roundworm Caenorhabditis elegans, a multicellular organism which had a 100 million base pairs in its genome.

The Race Begins

In 1998, the consortium was well-placed to begin the large-scale sequencing of the human genome. They had created a physical map which showed identifiable landmarks on chromosomes – such as the positions of disease-causing genes. The roundworm genome project had just been completed. Things were going well. Once again, Venter, who had founded a private company called Celera Genomics, burst into the scene. His team, he said, was poised to finish sequencing the human genome in a couple of years. The consortium immediately moved the deadline to 2003, two years ahead of the scheduled finish in 2005. Watson reminded the world that 2003 would be the 50th anniversary of the discovery of the double helix – so it was, in fact, an ideal date for the completion of the project.

Francis Collins, head of the N.I.H. consortium, planned to stick to their structured, map-based approach in which the DNA is broken into fragments, and the position of each fragment is mapped on the chromosome first. In Celera’s shotgun method, the genome was broken into millions of DNA fragments and pieced together in one go, without creating any map – the assembly called for sophisticated algorithms and greater computing power.

In the summer of 2000, at a gala event in the White House, both groups announced that they had arrived at a working draft of the human genome. The following year, the consortium and Celera would publish their results in the journals Nature, and Science respectively. The consortium had completed only 85 percent of the genome; Celera was not much further ahead. Both versions had known gaps and errors. One thing, however, was clear – the human genome had less than 25,000 protein-coding genes.

Celera moved on to other things, but the consortium kept pegging away at the draft. In 2003, on the day of the agreed-upon deadline, the human genome was declared complete once again. This version, too, was not error-free. The most complete genome sequence was still missing about eight percent of the genome. In 2019, a group called the Telomere-to-Telomere (T2T) consortium, decided to take on the challenge of arriving at the complete sequence.

Closing The Gaps

When the Human Genome Project began in the 1990s, sequencing machines could read only short stretches of DNA at a time – less than a thousand base pairs. So, a genome was broken into suitably small fragments and sequenced individually. A computer program would look for overlaps at the ends of the sequences and fit the fragments together in the right order to reconstitute the stretch.

But repeats in certain parts of the genome were so long that it was hard to figure out where the sequences fit. The problematic repeats occurred in biologically important regions: telomeres (regions at the ends of chromosomes) centromeres (typically reside in the middle of the chromosome) and short arms of five chromosomes, where centromeres are skewed toward one end. Transposable elements, the mysterious sequences that can move around the genome, are again full of repeats.

The source DNA used by the consortium also posed problems. The consortium had collected DNA from many anonymous individuals to get a mosaic human genome. Because of the hundreds and thousands of variations in sequences between individuals, some artificial gaps were created in the consortium’s genome. Celera’s DNA, it is reported, largely came from a single individual – that of its founder Venter. The T2T consortium, which used an unusual cell type that has DNA inherited only from one parent, sidestepped the variation issue altogether.

Long-read sequencers can read 10,000 bases accurately at a time — some can read 100,000 base pairs accurately-enough. Having reads that can span the length of the repetitive sequences made it easy to place the segments correctly in the genome. So, the T2T scientists arrived at a more complete sequence of the human genome and published their results in the Science issue dated April 1, 2022. The T2T-CHM13, as the new reference genome is called, represents the most complete, accurate human genome sequence there is yet –their DNA source did not have the Y chromosome. Finally, scientists can confidently say that the human genome has 3.05 billion base pairs. They have added, or fixed, more than 200 million base pairs in the reference genome. They estimate that our genome contains 19,969 protein-coding genes.

With the complete genome, researchers can finally study variation in DNA in individuals. Still, one reference genome does not convey the genomic diversity of the human species. We need many reference genomes–a pangenome. This monumental undertaking called The Human Pangenome is already taking place and is poised to redefine the future of genomic research and human health. Ultimately, the goal is that every person would be able to have their complete genome sequenced as part of their medical record – faster, cheaper and without using huge machines which take up a lot of room.

A much clearer, high-resolution picture of the genome has emerged now. Why is such a small part of the genome’s total length devoted to protein-coding? What is the function of the repeats? What do the non-repetitive, non-gene-coding parts of the genome do? Such questions promise to keep biologists occupied for a long time. The secret of life, which Watson and Crick thought they had stumbled upon when they discovered the doubled-helical structure of DNA, is yet to be deciphered fully. 

The post Probably A (Possibly C, G or T) appeared first on Vijee Venkatraman.

At the pediatrician’s clinic, a nurse told Viorica Marian, who is a native speaker of Romanian, to use only English with her American-born daughter. Speaking another language would “confuse” the child and hurt her long-term, the woman had said. This was a good decade ago. Even today, it is common advice for immigrants in the United States – it is also completely wrong.

In her new book, The Power of Language, Marian, a Moldavian American linguist, draws deep on research, some of it her own, to explain how language operates in our minds, and how we can harness the limitless power of languages to enrich our lives, as individuals and societies. She makes the convincing case that being bilingual, or better still, multilingual, can work wonders for the brain.

When bilingual persons use one language, she explains, the other language is active in their brains at the same time. As a result, the executive control system, whose job it is to keep us focused on what’s relevant, gets honed constantly. Just as exercise can change our bodies, this mental activity rewires the bilingual brain.

A buff executive control system gives bilinguals certain cognitive and social advantages even at a young age– they are good at multitasking, for instance. Later, they are able stave off the onset of Alzheimer’s, and other forms of dementia, by an average of five years compared to their monolingual peers with the same level of anatomical decay.

“If the brain is an engine, bilingualism may help to improve its mileage, allowing it to go farther on the same amount of fuel,” the author writes. The attention and aging benefits aren’t exclusive to people who were raised bilingual – they are also seen in people who learn a second language later in life. It is never too early or too late to start learning another language, the author emphasizes.

Because language and culture are intertwined, bilinguals may have different mindsets for each language. “Just as H2O can be a solid, a liquid, or a gas depending on temperature, a person can be a different version of themselves depending on which language they are using,” she writes.

If the idea that various versions of the self can coexist in a speaker of many languages, seems too romantic, consider plainer ramifications of bilingualism. When Mandarin–English bilinguals were asked to name a woman who succeeded despite physical handicaps, they were more likely to say Helen Keller when speaking English and Zhang Haidi when speaking Mandarin. They knew both answers, but what came to mind varied depending on the language spoken at any given time.

It is not just hard facts – even the recall of personal memories can vary depending upon the language. In a linguistically diverse nation like the U.S., the finding has implications for interviewing bilingual witnesses in legal cases.

Similarly, when providing psychotherapy to bilingual clients, therapists need to be aware that, in some instances, crucial early memories may be encoded in a client’s native languages and can only be retrieved in that language.

The majority of the world’s population is bilingual or multilingual. People who speak a language that is considered low-­prestige are well aware of the benefits of learning another language, ideally a dominant language that gives them access to the power dynamic of a globalized world and economy, Marian writes. It is the people in the First World who typically don’t see any advantage in being multilingual.

This book comes packed with interesting insights about the power of language even though it is written in the language of research papers. The codes we use to think, speak, and live, make for an endlessly fascinating topic. Chances are, you will download a language learning app, or make plans to sign up for an in-person language class once you’ve read this book.

The post The Power of Language appeared first on Vijee Venkatraman.

Coded Bias is a timely, thought-provoking documentary which follows Buolamwini’s journey to uncover racial and sexist bias in face-recognition software and other AI systems. Such technology is increasingly used to make important decisions, but many of the algorithms are a black box.

The documentary, which premiered at the Sundance Film Festival earlier this year, features a band of articulate scientists, scholars, and authors—primarily women of colour—doing most of the talking. This casting 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, due to the recognition of this problem of racial bias, there has been a backlash against the widespread use of 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.

People seem to have different experiences with the technology. The documentary shows a bemused pedestrian in London 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.

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.

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.

The director Shalini 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. 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.

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.

The post Coded Bias appeared first on Vijee Venkatraman.