Smart glasses are already sowing chaos in India: viral harassment and unchecked police surveillance

Camera-equipped smart glasses, led by Meta's models, are becoming a growing source of harm in India for people filmed without their consent, according to a report by MIT Technology Review. The case that opens the article is that of Shubnam, a 38-year-old transfeminine educator and graphic designer, filmed this year at a protest in Delhi against a bill that would have restricted legal recognition for trans people in India. A content creator known for 'rage-bait' videos approached her, asked her why she was wearing 'women's clothes' and recorded the entire exchange with a pair of Meta glasses shaped like ordinary Wayfarers. The video, edited as mockery, went viral on Instagram, X and YouTube, racked up millions of views, transphobic insults and AI-generated memes, and is still circulating months later. Shubnam says that single non-consensual video erased two decades of a hard-won personal journey, and she has avoided wearing a sari outside her neighborhood ever since.

The report places this case within a broader trend: smart glasses are one of the fastest-growing devices in India's wearables market, according to the consultancy IDC, and Reliance Jio —one of the country's largest telecom operators— plans to launch a rival model for under $105 before the end of the year, compared with the $420 that Meta's latest model costs in India. The central problem, notes researcher Janusz Świerczyński (Oxford Saïd Business School), is that these cameras do not look like cameras, so the people being filmed often do not even know that they are. In his view, the responsibility for spotting the recording falls increasingly on the person being recorded, who has to notice a small LED that lights up at the moment of recording, rather than on the manufacturer.

Meta maintains that its warning system works: the company says every pair of glasses has a capture LED that flashes when taking a photo or recording video, that it cannot be disabled, and that if anyone covers or damages it the camera is disabled. However, the report itself documents that ways to get around that warning already exist: tutorials that teach users how to cover or hide the LED while still recording, and a market for 'stealth mode' modifications —documented by Wired in March— that physically disable the light or remove it altogether. Świerczyński stresses that this kind of covert recording causes unequal harm: women and other marginalized groups suffer greater damage when discreet devices capture and spread images of them without consent, as happened to Shubnam.

Beyond viral 'pranks' between private individuals, the article describes a second, equally worrying front: the adoption of these glasses as a police surveillance tool. In June, when thousands of young people demonstrated in India over problems in the education system, Delhi police used Meta smart glasses to film the protesters. A petition filed with the Delhi High Court by Aishe Ghosh, former president of the Jawaharlal Nehru University student union, alleged that police had filmed the protesters continuously for weeks, even while they were eating or resting, and had threatened to send that footage to their parents and their schools. In July, India's solicitor general had the case dismissed in court, calling it 'luxury litigation'; instead of investigating the students' complaints, police opened ten criminal cases against the protesters themselves for offenses such as rioting, assaulting public officials and damaging public property. Several protesters, speaking on condition of anonymity, say they never saw the LED lit on the police glasses during the protests.

This combination —discreet cameras plus a police apparatus willing to use them to surveil rather than protect— reduces the chances of strict regulation of these technologies being imposed in India, according to the report. Sameer Parmar, a counselor at Meri Trustline (a safety partner that works with Meta and YouTube to support victims of harmful online content), explains that the hardware is improving faster than the social understanding of consent in a country where being in a public space is usually read as implicit consent to be photographed or filmed. According to Parmar, his helpline has recorded a marked increase in cases involving covert recording and non-consensual footage, including recordings in medical settings and private encounters, and he warns that those with the least power to object are the most vulnerable to this kind of harassment or filmed 'pranks'.

Lawyer Apar Gupta, founder of the Internet Freedom Foundation, points to an underlying legal gap: India's laws against voyeurism and the distribution of intimate images were designed with more recognizable forms of abuse in mind —filming someone undressing, using the bathroom or during a sexual act— so they offer little protection when covert recording happens in an ordinary public setting and is then used to humiliate the person. In addition, Gupta says, platform rules provide another route for taking down harmful content, but it is not an immediate solution: under India's current rules, platforms can take up to 36 hours to act on certain complaints, long enough for a video to be downloaded, copied and redistributed on other sites.

The article closes with a look at a future that compounds the problem: according to the Financial Times, some internal prototypes of new Meta glasses would include AI features capable of continuously photographing and listening to the wearer's surroundings, allowing them to recall later what was seen or heard, without those features triggering the external light indicator, which would make it impossible for third parties to know when information is being collected about them. Meta has said it will keep the LED for active capture —taking photos or recording video— but not for the AI features that interpret the surroundings, arguing that a constantly blinking light could end up being normalized to the point where people stop noticing it. Law professor Woodrow Hartzog (Boston University) warns that this logic risks locking society into a perpetual race to catch up with tech companies able to release these tools faster than society can grasp their threat, so that by the time it finally reacts, normalization has already taken place. For Shubnam, the underlying danger is not just getting used to being recorded, but getting used to versions of yourself circulating that you never chose to show: being understood through a narrative you neither composed nor consented to, she says, amounts to being left powerless over how others come to know you.