Over the last few weeks, the policing of women’s bodies has reached a new tipping point, as a draconian new ‘Hijab and Chastity Bill’ is being railroaded through parliament behind closed doors, prompting fresh fears about how new Artificial Intelligence and facial recognition technology is being used to arrest, target, and punish women. Almost a year on from the Woman, Life, Freedom uprising, the situation is deteriorating in Iran.
These new developments signal longstanding concerns ARTICLE 19 has been raising about the human rights harms caused by the development of facial recognition technology, which are especially horrifying in the hands of a regime that does not shy away from what the UN Special Rapporteur on the human rights situation on Iran has termed the Islamic Republic’s ‘crimes against humanity’.
ARTICLE 19 reiterates its long-standing position on the use of facial recognition technologies:
- ARTICLE 19 reasserts that a moratorium is needed on the development of facial recognition and biometric technologies given their fundamental inconsistencies with international human rights standards.
- Companies involved in the development of this technology must be held accountable by states and international bodies.
Below Azam Jangravi details these concerning new developments. Azam Jangravi is an Information Security Analyst and longtime women’s rights activist. She is known as one of the ‘Girls of Revolution Street’ – one of the first women to take her hijab off in protest on an electricity box on Revolution Street in 2018.
Since the Woman, Life, Freedom movement – the uprising following the death in custody of Mahsa Jhina Amini in September 2022 – began, the government of the Islamic Republic of Iran has been exerting pressure and intimidation on society.
Women in particular have been at the forefront of the movement, actively advocating against the injustices and discriminatory laws targeting women, most fervently against the mandatory hijab and Islamic dress code imposed on women. In response, the Islamic Republic of Iran ramped up efforts aimed at identifying and punishing individuals who do not adhere to the hijab requirement, enacting stringent laws, pressure and even deploying surveillance technology.
Over the past weeks and months, a surge of reports has emerged detailing the closure of commercial establishments, offices, and companies following the identification of women not adhering to hijab regulations within their premises. Notable cases include the sealing of the Digikala online store, the Azki insurance comparison website, and the Taghcheh online store. These closures were triggered by the dissemination of images featuring their female employees not wearing hijab attire on social media.
New developments and statements from authorities about policing women
Ahmadreza Radan, the Commander-in-Chief of the Islamic Republic of Iran, emphasised this stance during a statement in June 2023, stating that the expansion of surveillance coverage through Faraja cameras (the name used for CCTV cameras) is aimed at bolstering security from all angles and effectively drawing a line for women who reject the mandatory hijab.
These pronouncements gain significance when considered in the context of ongoing discussions surrounding the new bill related to the enforcement of modesty and hijab (known as the ‘Hijab and Chastity Bill’). Collectively, these efforts underscore the determination of the Islamic Republic to counter a national popular uprising for individuals’ right to choose whether or not to adhere to the veil with unabashed repression and further erosion of civil and political rights.
To further instil fear, regular output by state-affiliated media has sought to demonstrate the authorities’ tools to penalise women not wearing hijab. In June, Fars News Agency, closely associated with the Islamic Revolutionary Guard Corps, released a video depicting men telling women to wear headscarves and cover up. The filmed footage showed supposed facial recognition technology being used, whereby certain faces would lead the filmmakers to pull up women’s identity photos.
While it’s not clear what technology the Fars broadcast was using, or even if the matching of faces to identity photos were accurate, claims of using facial recognition technology to police women’s hijab and chastity in the streets have predated the Woman, Life, Freedom movement. In August 2022 Seyyed Mohammad Saleh Hashemi Golpayegani, the secretary of the Headquarters for the Promotion of Virtue and Prevention of Vice, a state-run body in charge of enforcing the mandatory hijab, explained in an interview: ‘Should an individual choose not to adhere to hijab while in public spaces such as the subway, surveillance cameras installed within the subway will capture an image of the person, subsequently resulting in a significant fine.’
Expanding on this notion, the Chairman of the Committee of Councils within the Islamic Council referenced the Minister of Interior’s announcement, stating: ‘Following an initial warning via text message, individuals who persist in disregarding the hijab requirement within public service venues will be denied access to services. If such individuals continue to resist compliance, the matter will escalate to judicial authorities for further inquiry.’
Ahmadreza Radan, the commander in chief of the police force, stated: ‘The implementation of the strategy involving surveillance cameras to address non-compliance with hijab is currently underway. Individuals who breach the hijab requirement are swiftly identified through this system. Initially, a first text message serves as a warning, after which subsequent messages involve coordination between judicial and police entities.’
Radan further elaborated, ‘Within the realm of public security and the discourse surrounding hijab, we have established a framework consisting of four supervisors. Each supervisor oversees a distinct approach to addressing instances of non-compliance with hijab. To exemplify, supervisor four is primarily active within the realm of cyberspace. This role, known as “Nazer Chahar,” is designated for individuals who identify cases of hijab non-observance online or those who advocate against its adherence. Subsequently, legal actions have been initiated against certain individuals, and requisite measures are being taken to address these situations.’
Under current plans for the Headquarters for the Promotion of Virtue and Prevention of Vice to approve the ‘Chastity and Hijab Bill’, individuals not adhering to the hijab will be categorised as offenders rather than criminals. This initiative outlines various levels of offences for individuals not wearing proper hijab. As stated by the head of this unit last year, these offenders are expected to be identified using surveillance cameras equipped with facial recognition technology, especially in public locations such as the subway.
However, what exactly is the capacity of artificial intelligence to discern individuals’ facial features, and how does this mechanism operate?
Facial recognition entails the recognition and intricate analysis of distinct facial attributes. Throughout this process, facial imagery is translated into digital data, subsequently cross-referenced against diverse databases to establish its identity. This sophisticated technology finds application in diverse scenarios, including unlocking smartphones, enhancing border control measures, fortifying system security, mitigating criminal activities, and spanning across a multitude of other domains.
Assessing the Islamic Republic’s assertions on Artificial Intelligence: Insights from experts
In response to assertions made by prominent figures within the Islamic Republic regarding the purported utilisation of artificial intelligence for identifying individuals who defy mandatory hijab norms, a number of technology experts and advocates have offered their perspectives on the matter. Notable among these voices is Ahmed Ahmadian, a technology authority based in California. In a dialogue with BBC Farsi, Ahmadian remarked, ‘The prospect of employing cameras to monitor hijab adherence carries a stronger psychological impact, aiming to instil a sense of fear. It appears that the Iranian government may not currently possess such advanced capabilities, but the Islamic Republic is rapidly advancing to acquire them.’
What adds intrigue to this scenario is that, subsequent to the proclamation of this assertion, Azari Jahromi, the Minister of Communication under the Hassan Rouhani administration, expressed reservations concerning the efficacy of this technological endeavour, deeming it a considerable misstep. Jahromi criticised the system, saying: ‘The vulnerabilities inherent in this system will inevitably come to light, possibly leading to a decline in its effectiveness. It’s conceivable that as of tomorrow, the promotion of counter-measures such as facial make-up designed to thwart surveillance could gain significant traction.’
Beyond the former minister’s worries over counter-measures, there are obvious questions over what exactly the Islamic Republic’s technological capabilities are for identifying both individuals and their type of dress for these chastity and dress code infringements. Do the existing hardware and software work together to deploy such surveillance? Questions of exactly what models and databases authorities will use for identification are unclear, as are the capabilities of the hardware to work with this software.
According to the Headquarters for the Promotion of Virtue and Prevention of Vice, the Islamic Republic’s database for identification predominantly comprises images from national identification cards. This raises questions over whether this is sufficient data for identification given differences in quality of images captured on street surveillance or mobile phones when compared to photos in identification cards. This highlights the danger of imprecise matches and wrongful persecution and fines.
Precedent of past technology projects
The roll-out of facial recognition and surveillance technology to police women’s dress is not the first digital repression project the Islamic Republic has announced. Hassan Rouhani’s presidency was marked by many projects to make internet controls efficient, including attempts at rolling out the ‘smart filtering’ project. This initiative aimed to ‘intelligently’ identify and censor specific content on the internet without blocking entire platforms. Most famously, authorities attempted to enact ‘smart filtering’ on Instagram. While not wanting to censor the entire platform, they wanted to ensure that pages and posts that went against Islamic morality and laws would remain inaccessible. Their success on Instagram was quickly dismantled once Instagram’s parent company Facebook (now Meta) initiated encryption on its platform in May 2015. On 21 October, 2015, Morteza Barari, the Ministry of Communications and Information Technology’s Deputy for Government, Parliament, and Provincial Affairs, disclosed, ‘an expenditure of 200 billion tomans (59 Million US dollars) was allocated for smart filtering in the previous year.’
There is no transparent account of how this money was used or why. The project has largely been deemed a failure, and an example of mismanagement in state funds. Even hardline parliamentarians have used the failure of the project as a reason to pursue the highly draconian ‘User Protection Bill’ in its place.
Who are the vendors behind Iran’s surveillance technologies?
The Islamic Republic has a known history of using surveillance technology to aid repression against its population. Human rights groups have long been documenting the uses of these technologies against activists. The most famous case is the surveillance and arrest of activist Isa Saharkhiz in 2009 using the Lawful Interception Management System (LIMS) that Nokia-Siemens sold to the regime. After a resolution against the company was issued by the European Parliament on 10 February, 2010, and a lawsuit was brought against Nokia-Siemens by the Saharkhiz family, the company stopped selling surveillance equipment to Iran and slowly stopped its business in the country.
In the face of oversight and sanctions placed on European counterparts, it’s generally known that China is the greatest exporter of technology the regime uses to aid digital repression.
Tiandy technologies has come under attention as one prominent exporter of surveillance equipment to Iran. Evidence has emerged indicating the sale of equipment from Tiandy to the Revolutionary Guards (IRGC), the powerful paramilitary organisation central to many of the human rights crimes committed against protesters and activists, and listed as a terrorist organisation by the United States.
The United States has sanctioned Tiandy in light of its role selling such repressive technologies to Iran in the midst of the popular uprising, which a United Nations expert has noted has been met by a level of state repression that possibly amounts to ‘crimes against humanity‘. These technologies of digital repression are thought to further facilitate these crimes. The United States has since been exploring the option of levelling more sanctions on the company.
Another prominent player suspected of selling digital repression technology to Iran is Huawei. Over the preceding years, this enterprise has faced accusations from the United States, specifically for providing equipment designed to identify protesters within Iran, an accusation that has attracted significant attention.
Reports have also emerged detailing the Iranian government’s deployment of street surveillance cameras crafted by manufacturers such as Bosch, originating from the Netherlands and Sweden to oversee protests and identify civilian demonstrators. As per Bosch’s assertion, these traffic monitoring cameras were supplied to Iran between the years 2016 and 2018. However, Bosch claims the cameras provided lacked comprehensive facial recognition capabilities, and could not have been used in repressing individuals.
An in-depth examination in the same report reveals that, through the utilisation of production software resources purchased from the Danish firm Milestone Systems in 2019, Iran was able to increase its capabilities for facial recognition and identification.
In this context, in June 2022, the hacker group Ghiyam Ta Sarnegouni hacked the websites and traffic camera systems of Tehran Municipality. They released videos on their social networks, asserting the utilisation of Milestone technology in the surveillance cameras positioned at intersections and highways in Tehran.
Human rights groups sound the alarm on surveillance cameras policing women
Amnesty International’s Secretary General Agnès Callamard has sounded the alarm on this new trend in policing women despite traditionally so-called morality policing having changed since the start of the uprising: ‘Today’s crackdown is intensified by mass surveillance technologies capable of identifying unveiled women in their cars and pedestrian spaces.’
Recent research published by Amnesty International documents numerous cases where women have been notified of ‘hijab infractions’ through these technologies. Upon crossing intersections or disembarking from vehicles, women receive prompt text messages indicating the capture of images showcasing non-compliance with hijab rules.
This report further cites the spokesperson of the traffic police boasting of 1 million SMS warnings to women captured unveiled in their cars since 15 April, 2023; 133,174 SMS messages requiring the immobilisation of vehicles for a specific duration; confiscations of 2,000 cars, and referrals of more than 4,000 ‘repeat offenders” to the judiciary across the country. The spokesperson added that 108,211 reports on the enforcement of compulsory veiling laws had been gathered about the commission of ‘offences’ within businesses and that 300 ‘offenders’ had been identified and referred to the judiciary.
The so-called ‘Chastity and Hijab Bill’
Iran’s parliament is currently considering a draft of the ‘Chastity and Hijab’ bill. Proposed and written by the Legislative and Judicial Committee of the Islamic Council, this 70-article bill encompasses a range of provisions that have stirred up debates due to their potential implications, particularly regarding aggressive new strategies to identify transgressors and the introduction of new offences.
This legislative proposal contains provisions for multifaceted penalties designated for violations of hijab protocols. While the Islamic Penal Code contains provisions criminalising not wearing mandatory hijab or promoting dissidence against it, the bill goes a step further, calling it ‘nudity’ for a woman to appear in public without hijab, thereby increasing its criminalisation and penalties.
As outlined, individuals who engage in acts such as disparaging the core tenets of hijab, whether in public spaces – both online and offline – or advocating for immodesty, indecency, non-adherence to hijab, or inappropriate attire, will face fines. Subsequently, upon the decree of the judicial authority, those found guilty may be subject to travel restrictions and banned from using the internet from between 6 months to 2 years. Offenders could face fines of up to 8 million tomans ($150), with the fines being doubled if not paid within a month, lose their jobs and be banned from social media activities for up to one year.
If this bill is approved and ratified, for the first time in the Islamic Republic, an offender’s personal assets will be confiscated by law enforcement without any judicial proceedings. The bill proposes to directly deduct penalties from an offender’s bank accounts and confiscate passports.
In a worrying move on 12 August, 2023, the parliament voted to pass Article 85 in relation to the progress and implementation of the bill. Article 85 means the decision over the implementation of the bill is left to a small group of MPs working behind closed doors as opposed to a general vote amongst all parliamentarians. This curbs accountability and threatens transparency in the final draft of the bill. After the small group finalises the draft, the bill will be referred to a parliament open session. During the session, it will be determined how long the bill should be implemented ‘experimentally’, and then it will be sent to the powerful (and unelected) Guardian Council to be ratified.
Notably, Article 85 has also been invoked for the controversial ‘User Protection Bill’ known as ‘Tarhe-Sianat’. Regarding the ‘Tarhe-Sianat’ initiative, certain Members of Parliament held the perspective that ‘the Supreme Council of Cyberspace circumvented parliamentary approval in sanctioning the “Tarhe-Sianat’ bill”.’
The actions taken in the lead-up to the anniversary of Mahsa Jhina Amini’s death underscore the government’s concerted efforts to deter the resurgence of protests. Employing an array of intimidation tactics and generating a climate of fear, including the utilisation of facial recognition technology and dispatching text messages regarding hijab compliance are indicative of these intentions. The aim is to curtail public presence on the streets.
Yet the pivotal question that lingers pertains to the enduring gains these strategies might confer upon the Islamic Republic. The response to this question will unfold over time, offering insights into the long-term implications of these policies.