Image of map and computer language over binary codes on server racks in server room.

Global Digital Policy Roundup: May 2026

Abhijeet Dipke, center, founder of the Cockroach Janta Party addresses supporters during a protest demonstration in New Delhi, India, Saturday, June 6, 2026. (AP Photo/Manish Swarup)

After Test Data Scandal, India’s Youth Stake Their Claim on the Digital Square

A blur effect added to Leonardo da Vinci’s Mona Lisa

AI Can Rebuild Blurred Faces, So How Do We Protect People Now?

Close-up of Keir Starmer, wearing a dark suit, white shirt, and tie, looking directly at the camera.

Prime Minister Sir Keir Starmer delivers opening remarks ahead of a meeting with senior figures from TikTok, X, Meta and other social media giants at 10 Downing Street, Westminster. Picture date: Thursday April 16, 2026. 84281335 (Press Association via AP Images)

Could a UK Teen Social Media Ban Work Without Expanding Surveillance?

The image shows a hand holding a smartphone in front of a computer screen with war and other words projected across.  

Hidden Spyware Use in War May Be More Widespread Than We Realize

How OpenAI, Google, and Anthropic Plan to Handle the 2026 US Midterms

Panelists at a session titled “Does AI Change How Change Happens?” at Netroots Nation in Philadelphia on June 6, 2026. Dean Jackson/Tech Policy Press.

At Netroots Nation, Progressives Divided on AI

LONDON, UK—NOVEMBER 15, 2025: Panorama of Google logo sign on door of HQ building in Kings Cross.

The UK Just Said No to Google's AI News Grab. Who's Next?

Three colorful GPUs with their packaging cleanly removed laying on a white surface.

The China Chip Strategy That Is Backfiring on America

A red domino inserted amongst a bunch of neutrally colored dominos, meant to symbolize the intervention of a red team in AI systems.

When AI Fails, What Actually Failed? The Distinction AI Governance Keeps Missing

A youth holds a Brazilian flag as residents prepare street decorations ahead of the 2026 FIFA World Cup, in Brasilia, Brazil, Sunday, May 17, 2026. (AP Photo/Eraldo Peres)

Brazil Moves to Increase Citizens' Security Online

Stop the FISA Fearmongering and Allow Fair Votes on Reforms

A photo of four EU flags billowing in the wind with a blurred office building behind them.  

The DSA’s Sleeping Giant is Waking Up

We Need to Know More About How AI is Affecting Mental Health

Teen student wearing headphones using laptop in city urban location searching information, online learning, watching webinar class or having video chat outdoors.

Teens are Relying on AI for Emotional Support. Policymakers Must Respond

Randi Weingarten, President, AFT, speaking at the 2026 J Street National Convention held at the Hyatt Regency Hotel in Washington, D.C. (Photo by Michael Brochstein/Sipa USA)(Sipa via AP Images)

In Pivot, Teachers Union Pitches 'Devices Down, Eyes Up, Hands-On' Plan

In this photo illustration, the Palantir logo is seen displayed on a smartphone. (Sipa via AP Images)


Why Palantir’s UK Health Data System Matters Beyond Surveillance Fears

Mountain View, California: The Google manager responsible for the search engine, Liz Reid, presents a search field improved with artificial intelligence at the Google I/O developer conference on May 19, 2026. Photo by: Andrej Sokolow/picture-alliance/dpa/AP Images

Google’s ‘New Era for AI Search’ May Threaten Democracy

The dome of the US Capitol building. Justin Hendrix/Tech Policy Press

House Subcommittee on Cybersecurity and Infrastructure Protection Hosts Hearing on AI Security

June 13, 2024 - The United States Capitol. Justin Hendrix/Tech Policy Press

House Subcommittee on Commerce, Manufacturing, and Trade Hosts Hearing on SECURE Data Act

Vas Panagiotopoulos is a fellow at Tech Policy Press.

Mobile phone spyware has been used on several occasions during wartime in recent years. Originally marketed as a tool for combating terrorism and serious crime, spyware has increasingly appeared in military and intelligence operations around the world.

Spyware can be used for targeting and gathering battlefield intelligence, military espionage and counterintelligence, as well as psychological and deception operations.

States value these capabilities because they can provide a distinct strategic advantage over adversaries that do not possess them. However, these military-grade technologies can cause serious harm to human rights and create significant counterintelligence risks.

Spyware is difficult to detect and document, especially during armed conflict, where there may be no opportunity to examine a device and evidence can easily be lost.

There are several publicly known cases of spyware being deployed in armed conflicts in recent years, including in Iran, Gaza, Ukraine, Nagorno-Karabakh and Syria. Yet the cases documented by researchers to date suggest that what we currently know is only the tip of the iceberg.

What follows is a non-exhaustive list of publicly documented cases involving the deployment of spyware during armed conflicts in recent years:

Targeting and battlefield intelligence collection

Spyware can be used to identify, track, and validate military targets by collecting data from compromised devices and networks, as well as to collect battlefield intelligence — gathering information on enemy communications, locations, movements, plans, and capabilities to support operational decision-making and targeting.

For example, on 16 June 2025, Israeli jets bombarded a bunker where an emergency meeting was taking place, attended by key Iranian figures, including President Masoud Pezeshkian, the heads of the judiciary and the intelligence ministry, and senior military commanders. None of the officials carried mobile phones, aware that Israeli intelligence could track them. Yet, according to an 

, Israeli forces were led to the meeting by hacking the phones of bodyguards who had accompanied the Iranian leaders to the site and were waiting outside. The piece that cites both Israeli and Iranian sources does not provide more details on the method by which these phones were compromised; however, Israelis are well-known for several such offensive cyber capabilities – including in-house ones, as well as commercial tools such as those by NSO Group and Paragon Solutions.

Another more concrete example of the deployment of commercial spyware for targeting and battlefield intelligence collection can be seen in Gaza. In the aftermath of October 7, Axios 

 that several Israeli agencies were likely using NSO Group’s Pegasus spyware. “Pegasus can be used to tap into cell phone signals to assess who was on the ground during Hamas' surprise attack and movement of those cell signals before and after the attack,” noted the piece, quoting an anonymous source with direct knowledge of NSO's operations.

The source said that NSO had also established a so-called “war room,” bringing together other similar companies and former NSO employees to track and unlock phones belonging to people who have been murdered or gone missing, as well as those of suspected terrorists.

 in 2016 the battlefield deployment of the iOS and Android spyware X-Agent against Ukrainian artillery units by the Russian cyber-espionage group Fancy Bear, likely linked to the Russian military intelligence agency (GRU). The malware 

 a device’s microphone to record audio and collect, among other data, text messages, contact lists, photos, and geolocation information.

Military espionage and counterintelligence

Spyware can also be used for military espionage to covertly obtain sensitive military, political, diplomatic, scientific, or industrial information from adversaries, as well as for Counterintelligence to monitor, detect, or disrupt hostile intelligence activities and identify compromised personnel.

For example, a new US Defense Intelligence Agency report 

 by incidents in which American defense personnel in Israel discovered spyware secretly installed on their phones to intercept their communications, highlighting the serious counterintelligence risks associated with such technology.

 revealed that at least twelve Armenian public figures and officials, including journalists and human rights defenders, were targeted with NSO Group’s Pegasus spyware amid conflict in Nagorno-Karabakh in 2020 – 2022. This came after the Pegasus Project had 

revealed more than 1,000 Azerbaijani numbers

 were potential NSO targets and was the first documented evidence of the use of Pegasus spyware in an international war context.

The first cluster of civil society Pegasus infections in Armenia emerged against the backdrop of the 

2020 Nagorno-Karabakh war with Azerbaijan

, its aftermath and ceasefire, which triggered a domestic political crisis, mass protests, an alleged coup attempt, and ultimately 

Prime Minister Nikol Pashinyan’s resignation

Assad’s army officers were hacked through an app that deployed SpyMax, a widely used Android surveillance tool, New Lines Magazine 

 in 2025. The malware enabled keylogging to steal passwords and intercept text messages, extracted confidential files, photos, and call logs, and accessed the camera and microphone for real-time surveillance of its victims. The piece notes that it is difficult to determine exactly how many phones were compromised in the attack, but the number is likely “in the thousands,” adding that the unique element of this “primitive but devastating” phishing attack seemed to have been “focused on compromising an entire military institution.”

Spyware can also be used for psychological and information operations to enable influence campaigns on target audiences, as well as for military deception operations to manipulate adversary perceptions, decision-making, and situational awareness.

 in March that Iran had hacked Israelis using Android phones as they were fleeing an Iranian missile strike, an operation that required sophisticated coordination. They received a text message with a link purporting to provide real-time information about bomb shelters. However, the link was malicious and installed spyware, giving hackers access to the device’s camera, location, and all its data. The piece notes that such disinformation campaigns are not designed to kill, but to “to spy, steal and frighten.”

, the CIA used NSO Group’s Pegasus spyware to carry out a deception campaign in Iran as part of efforts to retrieve the second of 

 in early April. The report claimed that the American intelligence agency used Pegasus to send messages to Iranian leadership and Islamic Revolutionary Guard Corps operatives, falsely claiming that the downed US airman had already been found.

It’s worth pointing out that most of these cases are impossible to independently verify. John Scott-Railton, a Senior Researcher at The Citizen Lab who has researched Pegasus since 2016, 

 report’s claim: “No clear sourcing. No official confirmation. Just dropped in there.”

Spyware Use in Conflict: Which laws apply?

Legal oversight largely disappears in wartime. Spyware deployment during conflict often occurs on unclear legal grounds, with key safeguards — such as prior judicial authorization, and requirements of necessity, proportionality, and post-surveillance oversight — frequently completely absent. This creates conditions ripe for abuse.

So, are there any legal safeguards when spyware is used during war? In conflict, it is 

 that applies, which forbids the deliberate or indiscriminate targeting of civilians, humanitarian personnel, and other protected parties.

“Spyware cannot be used to harass, intimidate, and leak data of civilians, war prisoners, and other protected individuals,” Natalia Krapiva, Senior Tech-Legal Counsel at Access Now, told Tech Policy Press

, although the use of spyware is not a crime in itself, it may be part of the conduct that forms the basis of war crimes, crimes against humanity, or genocide.

“For example, if it is used to target and persecute civilians and other protected populations, provided that the requirements of intent, gravity, and other requirements are met for the underlying crimes,” continues Krapiva.

The use of spyware may also constitute an offense against the administration of justice if spyware is used to interfere with witnesses, evidence, or otherwise the court proceedings in criminal trials.

 and domestic law — including the right to privacy and freedom of expression that spyware may violate — usually also apply in conflict, unless a country derogates from these rights under the established procedure, in which case some laws may be temporarily suspended for as long as the public emergency situation is in effect.

“For example, Ukraine has derogated from certain obligations under the European Convention on Human Rights and the International Covenant on Civil and Political Rights (ICCPR),” says Krapiva.

“However, some rights like the right to life, the right to be protected from torture or slavery, are non-derogable, so even in conflict you cannot use spyware to torture people or to kill civilians.”

There is very little legal precedent concerning the use of spyware during armed conflict. Even in peacetime, spyware is often difficult to detect and document; in conflict settings, these challenges are amplified. Access to compromised devices may be impossible, forensic investigations may be delayed or infeasible, and critical evidence can be destroyed, lost, or rendered inaccessible.

The documented cases identified to date span multiple conflicts, governments, and spyware technologies. Researchers caution, however, that these incidents likely represent only a fraction of the true extent of wartime spyware use. Because spyware is specifically designed to evade detection, and because armed conflict creates significant obstacles to forensic investigation, the full scale of its deployment may never be known.

“From the cases that we do know and have documented, we suspect that there is more spyware used in conflict than we know about,” says Krapiva.

Jade-Ruyu Yan is a UK Reporting Fellow at Tech Policy Press and openDemocracy.

There is a social media ban for children looming on the horizon in the United Kingdom. Many are concerned that the government is not focusing on the unintended consequences.

The idea of a ban has been brewing in the UK for some time, in response to high-profile cases involving social media harms to children. Although the UK has 

 its wide-ranging Online Safety Act from 2023, there are “ongoing concerns about children’s online safety,” 

“It’s very difficult to know what the actual ban would look like,” said Alan Woodward, professor of computer science and cybersecurity at the University of Surrey. The details have not been announced yet, although 

specific restrictions have been discussed

The government launched a national consultation in March of this year that dealt with how to improve children’s relationships with phones and social media, and asked 

whether the UK should implement the same social media ban

The consultation, which closed last month and received over 116,000 responses, asked for responses to various types of restrictions, 

 the UK should restrict livestreaming and location-sharing, or require tech companies to disable features including infinite scrolling and autoplay. Other measures include how to tighten age verification rules, such as face age estimation technology or digital IDs, as well as guidance around mobile phones in schools and simpler parental controls.

UK Prime Minister Keir Starmer also announced

 that the government will require hardware manufacturers including Apple and Google to activate or install software on personal devices to detect and block explicit images, which could only be bypassed by verifying that you’re an adult.

If tech companies do not comply with this within three months, the government has said, they will face a legal requirement to add these locks to any technology sold in the UK and could face fines and criminal liability — the specifics of which are unclear, although legislation “could cover operating system providers and others in the supply chain, such as retailers” if they don’t comply.

 that it should not implement a ban, arguing that it would “impose disproportionate compliance burdens on American companies” and affect freedom of speech.

 she isn’t concerned “in the slightest” by the US’ warning. “I think companies will continue investing in Britain,” she said. “My focus is on what is right for British parents and British families.”

How would any restrictions on teen access work in practice?

 type of nudity-blocking software, although 

 for people using iPhones and iPads in the UK for specific apps and services.

also expected to deal with limits on AI chatbots

, which were not initially covered by the UK’s Online Safety Act and became a focus after Elon Musk’s AI chatbot Grok was 

used to create millions of sexually explicit images

“I don’t think the government really knows where they’re going with a lot of this stuff,” said Jasleen Chaggar, senior legal officer for British civil liberties and privacy campaigning organization Big Brother Watch, speaking about the prospect of a social media ban 

. “They want an outcome, but they haven’t actually figured out the technical solution to get there. So it’s like, ‘tech companies, you need to figure this out or we’re just going to outsource this to ID companies and get them to figure out the details.’”

When asked about the status of the ban, a spokesperson for the Department for Science, Innovation and Technology said in an emailed response, “We don't comment on speculation about future announcements.”

"We've been clear we are determined to act quickly, but we will do so in a way that is effective, enforceable and genuinely keeps children safe. As we've previously said, we will set out the Government's response by the summer, and important[ly] we already have the powers to act within months rather than years."

Conversation around a ban has been gaining traction, with 

 at the start of the year. There has been growing concern around online safety, with the 

government proposing earlier this year to give ministers unfettered power

launched a six-week trial with groups of children

 that tested their responses to four different types of ban of apps including Snapchat, Instagram and TikTok, with the results to be used as the “evidence we need to take the next steps,” said UK Technology Secretary Liz Kendall at the time. These included a blanket ban, daily time limits of one hour, a curfew blocking access from night to morning, and unrestricted use of social media.

 Australia’s social media ban, rolled out in December last year, as inspiration for the UK’s plan. Australia’s ban 

 a platform that enables engagement between two or more users, and allows users to post content.

"It is right that the government should tackle Big Tech's failure to prioritize child safety over profit, but we are concerned that the complex task of safely changing the digital landscape is being done on the hoof,” said Big Brother Watch’s Coulson.

 for political reasons. Critics have speculated that an embattled Starmer, who 

, including his safeguarding minister Jess Phillips, who cited slow progress on online safety. They joined 

A large coalition of critics has warned against the unintended consequences of this type of ban, in Australia and now in the UK.

Concerns focus in part on surveillance: organizations have warned that requiring platforms to verify the age of children requires them to verify everyone’s age, leading to a lack of online anonymity and the threat of data breaches.

 some form of ID online such as an ID or credit card check, or a facial scan, “sounds simple” and “has got widespread appeal,” said cybersecurity specialist Woodward. “You get a lot of political points for it.”

But in reality, “it means age-verifying all of us,” he said, as everyone will be required to submit identification online to show that they’re over 16. Alternatives include using AI to verify age, he said, but 

. He added that the discussions to put age verification on VPNs, which many have used to bypass age verification, are creating a “whack-a-mole situation.”

 people to verify their ages on websites that contain pornographic content.

"These policies create new risks, for adults and children alike,” said Big Brother Watch’s Coulson. He added that age verification introduces new data risks, as it’s not certain that platforms will delete peoples’ sensitive data.

Over the past year, for example, the platform 

 that over 70,000 user photos were potentially leaked, while users of the dating safety app Tea 

had their personal data and photos leaked

“The price of admission to the internet will now be ID checks,” he said. “Ending anonymity online is a step that authoritarian regimes can only dream of. We all deserve privacy but political dissidents, journalists, and victims of coercive control will find the need to out themselves especially terrifying."

When it comes to age verification, “people will say, ‘well I’ve done nothing wrong so I’ve got nothing to hide.’ And that is such a naive argument,” said Woodward. “Once you put the technology in place, time and again we have seen it being abused by govts–you get this scope creep.”

There are concerns, including from Scotland’s children’s commissioner in response to the UK’s consultation, that a 

ban could send children to more unregulated parts

 of the internet. Others are worried about cutting off the benefits social media provides young people, such as 

giving LGBTQIA+ youth a chance to access communities

The ban comes as countries including France, Greece, and Indonesia are 

considering and preparing for social media bans.

Critics have argued that this type of ban is hard to enforce: 

 61% of children and teenagers in Australia still have access to social media. There are 

 that it makes children ill-equipped to handle social media and the internet when they get older and do have access.

“We’re not honest enough about what these platforms are,” said Velisalava Hillman, a lecturer at Goldsmiths and founder of EDDS Institute, which researches and audits educational technology and AI systems used in schools. “They’re not about socializing or creating friendships–they’re moneymaking machines.”

If we “compare the opportunities and the current risks,” the risks far outweigh the positives and “are increasing in breadth and depth,” she said.

The idea of a ban is “a very narrow way of looking at this problem,” said Hillman. “The debate is always around the idea that children necessarily have to be on these platforms. They don’t have to, the world existed well before these social media platforms.”

“We await the government's wider proposals, but this plan is a retrograde step for everyone's privacy,” said Coulson. “What we need is a wholesale culture change around the age at which children are given devices, what kinds of devices, and how parents engage with existing tools."

 in crime science investigation television shows. The lead investigator leans over a forensic expert and says “Enhance the image,” and after a few clicks, a blurry face snaps into focus. For years, we used this clip in human rights documentation and verification training to mock Hollywood, stating that you can’t enhance what’s not there. We believed pixels don’t magically appear. However, that was years ago, and times have changed.

Last year, as new AI tools and capabilities emerged at breakneck speed, a question began to nag at us as human rights workers focused on visual research: are the blurs we rely on to protect people in photos and videos still doing their job? Or with enough time, tooling and incentive, could a government, a well-resourced actor, or even a motivated individual now reverse them?

Many human rights organizations and media outlets frequently use blurs to obscure the identities of people for their protection, to protect their security, privacy or dignity, or to reduce our readers’ exposure to distressing content. As a test, we took a published photograph of two children with their faces blurred and ran it through some commercially available AI “refocusing” tools, and facial structure re-emerged. Features sharpened. A plausible face appeared. Even if the reconstructed image did not reveal the children’s actual faces, it narrowed anonymity. It suggested age, gender, ethnicity, and resemblance. It moved the image closer to identity, and thus closer to harm. The question is not whether these tools are perfectly accurate today, but where they are heading.

Three side-by-side versions of Leonardo da Vinci's Mona Lisa, showing the same painting at different levels of facial clarity. Left: the original sharp portrait. Center: the same image with the face heavily blurred while the background stays in focus. Right: the face reconstructed/de-blurred, but with noticeably altered, younger and smoother features that differ from the original.

Unblurring test using the Mona Lisa. From left to right: the original image, the blurred image, and the image after unblurring.

This is no longer a hypothetical question. In January 2026, 

 on online users attempting to identify the ICE agent who fatally shot Renee Nicole. Using AI tools, they transformed a partially obscured image into a generated, reconstructed face, one that could potentially be compared against databases. The original photograph was never intended to reveal biometric data. AI turned it into something that could. The line between “non-identifying” and “identifiable” collapsed.

This is not new behavior. People have previously used open-source refocusing tools to identify Tinder users who “liked” their profiles, demonstrating both the feasibility of reversal and the existence of actors willing to attempt it. Media forensics expert Hany Farid shared a 

 with us from nearly 20 years ago, in which a serious criminal had posted photos of themselves online and used a spiral blur to hide their face. By simply reversing the blur with the same tool that created it, Photoshop, the individual was identified, leading to their arrest. Today’s AI shortcuts make that process faster and far more accessible. What has changed is scale, accessibility, and power—AI now amplifies these capabilities exponentially. For activists, witnesses, children, and bystanders, the stakes couldn’t be higher.

When we showed that crime scene clip in our training years ago, we were right, but only partially. The core problem is this: traditional blurring techniques—Gaussian blur, pixelation, soft focus—do not fully remove or obscure information. They degrade it. In photos and videos, blurring reduces spatial detail by averaging or obscuring pixel values. To humans, this loss of detail looks like disappearance: faces become indistinguishable, features seem unreadable. But to machines trained to recover a 

, the underlying information often remains present. It is distorted, but not erased.

Blurring was designed for human eyes, not machine vision. Modern computer vision systems are explicitly trained to operate under degraded conditions: low resolution, motion blur, extreme angles, partial occlusion, and noise. In the era of generative AI, blurring is increasingly ineffective as a protection measure. 

 for example, are designed to reconstruct images by progressively denoising random patterns. When blur or pixelation is applied, it functions as a predictable form of noise, one that these models are especially good at reversing or approximating. Similarly, recognition systems are trained to infer identity from incomplete cues and can be inverted or paired with generative models to produce clean, plausible reconstructions, including frontal portraits. What appears lost to humans may remain recoverable to an algorithm. For human rights researchers, continuing to rely on blurring techniques risks creating a dangerous illusion of safety for people whose lives may depend on anonymity. If we think of the blur as visual identity encryption, that encryption has now been cracked, or is about to be.

These capabilities are no longer experimental. They are commercial, widely accessible through 

, and improving rapidly. And this shift is already changing practice. In 2024, Human Rights Watch revised how we film people who require anonymity for safety reasons. Rather than relying on darkness, blurring, or post-production effects, our teams now avoid capturing identifiable visual data in the first place—filming people looking away from the camera, behind an object, or filming their shadow and carefully excluding identifiable features, landmarks, and locations.

Leonardo da Vinci’s Mona Lisa with her face obscured by a black square

Leonardo da Vinci’s Mona Lisa obscured by a black square

To understand what this means in practice, we spoke with leading experts and asked a simple question: how robust are our blurs, really?

Every expert gave the same answer. There are no assurances that a blur applied today cannot be reversed tomorrow, through a process officially called deconvolution. Technological progress makes guarantees impossible.

Their most consistent recommendation was also the most blunt: if you need to protect identity, cover it completely. Faces, bodies, tattoos, license plates, buildings—remove them with solid blocks of color. The “black square” approach.

The uncomfortable technical truth is simple: if identifying information remains in an image, it can potentially be recovered. The only reliable way to prevent reconstruction is removal. Black squares work. Full redaction works. Once data is gone, it is gone.

Simple. But not quite. Blurring exists for a reason. We blur to protect bystanders, children, perpetrators, we blur a face, a body, a tattoo, a street sign with the idea that we are obscuring identities, reducing harm and publishing evidence responsibly all while still communicating to our readers an idea of what the image shows. A blur preserves context, and a hint of the face behind the blur. It allows a reader to understand what is happening without drawing attention to the act of redaction itself. It is visually familiar. It signals care without overwhelming the viewer. It preserves emotion and narrative flow.

Replacing blurs with black squares would introduce a new visual language, one that may initially feel harsh or distracting, or worse, it may remove the humanity from the image. Perhaps audiences would adapt. Perhaps they would not. Once you start to notice “black squares,” it becomes apparent that they are slowly making an appearance, and it seems that we are not alone in asking these questions. Just look at the recent Jeffrey Epstein-related photos the US Department of Justice released at the end of January 2025—not a blur as a method of redaction in sight.

A grid of color snapshot photographs with several frames blacked out (redacted). Visible photos include: an older man with gray hair in a pink checked shirt outdoors on a sunny field; a man in an orange polo and gray jacket standing outdoors; a man wearing red aviation headphones seated in the cabin of a small private plane; a man in a gray "The Ranch Colorado" sweatshirt seated on a private jet with a woman resting her chin on her hand beside him; three people and a brown horse outside a long white stable building; a shirtless young person in white pants standing on a grassy lawn holding an object; and a pink-and-white Spanish-colonial-style building with arched arcades, balconies, and a landscaped garden courtyard. Multiple cells in the collage are solid black where images have been removed.

Screenshot of images from the Epstein emails released by the Department of Justice, showing black rectangles and squares being used to redact information. Screenshot taken from 

But total redaction comes at a cost. Visual evidence remains essential for documenting abuse, countering denial, and amplifying voices that would otherwise be ignored. Images matter because they convey humanity, emotion, context, and credibility. Even a blurred face can still give a human presence to the atrocities we report on. Replacing that with a black square or complete visual erasure risks weakening our ability to communicate harm, mobilize empathy, advocate effectively, and ultimately contribute to change.

Claude responded: Four side-by-side close-up crops of the Mona Lisa's face, illustrating an AI de-blurring or face-reconstruction process.Four side-by-side close-up crops of the Mona Lisa's face, illustrating an AI de-blurring or face-reconstruction process. From left to right: the original painting with the face sharp and intact; the same image with the face heavily blurred while the hair and background remain clear; an AI-reconstructed version showing a sharp but noticeably altered face with younger, smoother, more conventionally modern features that differ from da Vinci's original; and a final version where that reconstructed face has been blurred again, showing how the new (incorrect) features compare against the original blur.

From left to right: the original Mona Lisa, Mona Lisa with Gaussian blur applied, an AI-generated face superimposed on Mona Lisa, and the AI-generated face with Gaussian blur applied.

All the experts we consulted suggested a second, more experimental approach: using generative AI itself as a protective layer.

The idea is counterintuitive. Instead of blurring a face, first replace it with an AI-generated version of the same face and then blur that face. This would ensure that if the blur is later reversed, what emerges is not the original person but a fabricated stand-in. Returning to the encryption analogy, if the encryption has been cracked, change what you are encrypting.

We tested this approach on footage of a child crying in an ambulance, with a few clicks in Photoshop, we selected and masked the face area, entered a short text prompt, and used the built-in generative fill tool to synthesize a new face. When blurred, we got back an older man looking like he was screaming in pain.

This approach raises a series of uncomfortable questions. Would people inadvertently add a sensitive face to AI training data? How accurate would the AI face be, and if it were too close to the person we were trying to protect, could that resemblance expose them (or someone who looks similar) to mistaken identity, with serious real-world consequences? There are also deeper ethical concerns. How comfortable are human rights organizations using generative AI to edit evidence? What will the cost be for the integrity of human rights evidence and human rights organizations themselves, exposing them to 

 for using generative AI to edit evidence? And, for those needing to blur-on-the-go or without access to expensive tools like Photoshop, how would this solution work for small media outlets and civil society organizations?

Third solution—a really, really strong blur

Mona Lisa with a Gaussian blur applied to the face.

A third path was suggested: not blur alone, but blur as part of a layered anonymization pipeline. Instead of relying on a single visual transformation, this approach combines multiple degradations so that no single reversal technique is sufficient to reconstruct identity.

Experts emphasized that layering matters more than intensity. A single, heavy blur can still preserve structural signals that AI systems are trained to exploit. Multiple, overlapping transformations are far more effective.

In practice, this can include, in addition to replacing real faces with synthetic ones before blurring, 

 to permanently discard visual information; heavy compression to introduce irreversible artifacts; adding noise or randomness to disrupt statistical patterns; and expanding redaction beyond the face to hair, neck, shoulders, or bodies. For video, temporal techniques, such as variable blur that changes across space and time or frame degradation, can prevent identity cues from accumulating over time.

These techniques together raise the cost of reversal, forcing an adversary to overcome multiple interacting obstacles rather than applying a single “deblur” tool. Small implementation details: order, randomness, consistency across frames, can make the difference. This approach adds intention to the small details: slowing re-identification, increasing uncertainty, and adding friction in an environment in which perfect anonymization is impossible.

One answer we have settled on is intentionality: publishing visuals only when there is a clear and defensible reason to do so. Rather than blurring and republishing by default, we prioritize photos and videos with genuine evidentiary value, understanding that the solution is not less imagery but greater responsibility in its use.

We need to move from cosmetic blur to structural anonymization. Blur may still have a role, but pretending it is sufficient alone no longer protects potentially vulnerable sources and witnesses. We must stop treating blur as an ethical endpoint and start treating anonymization as a core safeguard that is designed, tested, and maintained with the same seriousness as data security.

We must also acknowledge a harder truth: there is no permanent solution, only risk reduction. In a rapidly evolving, adversarial technical environment, one-size-fits-all techniques are especially fragile, especially if these techniques are described publicly in detail. And the risk is not limited to faces. Identity leaks through bodies, posture, gait, scars, tattoos, clothing, background objects, architecture, signage, and location cues. Even when a face is blurred, an individual may remain identifiable, especially when images are combined with other data, revisited later with more powerful tools, or analyzed using multimodal AI systems that see patterns faster and more comprehensively than humans ever could. We could use AI tools to stress-test our anonymization by asking them what would help them de-anonymize a photo or video.

We need more resilient, layered, and adaptive approaches, ones that account for how both humans see and algorithms infer. These approaches should aim to remove identifying signals rather than merely obscure them, while still preserving the humanity of the people depicted. Such practices may require new visual norms; ways of preserving dignity without preserving identity. Above all, they demand clarity about risk, about limits, and about responsibility. Media platforms such as YouTube, which added built-in blurring tools following sustained engagement and advocacy by 

, should now also reckon with the risk of over-promising protection, as these tools were designed for an earlier technical reality and may no longer provide the safety users assume.

This is not about achieving perfect safety. As Hany Farid told us in an interview, 

. The goal is friction: to raise the cost of re-identification high enough that harm becomes unlikely within a realistic threat model. Anonymization cannot be “set and forget.” Techniques that seem adequate today may fail tomorrow. Standards must be reviewed, tested, and revised.

That raises difficult questions. Should previously published images be reassessed? How do we audit what is already online? Should organizations test their own imagery against reconstruction tools? How accessible will new protocols be globally? How transparent can protective methods be without empowering adversaries?

Avoiding these questions does not eliminate risk. It merely transfers it onto the people depicted. Blurring, as we have known it, is over. What comes next needs to be stronger, more honest, and worthy of the trust placed in it.

As the performance of generative AI models continues to improve significantly year-on-year, each new election season brings novel AI risks: AI-generated images are now all but impossible to discern from genuine ones; generated copy is less stilted and obvious; increasingly autonomous agents could potentially collapse the costs of running deceptive campaigns; and new model capabilities raise fears of cyber attacks on campaigns and election infrastructure.

That concern prompted Sen. Mark Warner (D-Va.) to issue a 

 to tech companies in March of this year recommending a series of measures intended to prevent the creation and dissemination of deceptive election-related content and to blunt its credibility ahead of the 2026 US midterm elections. (Warner sent similar 

 requesting information about what measures companies were taking ahead of the 2024 elections.)

At the same time, AI companies are also on the defensive against 

Latest

Podcast

Recent

Our content delivered to your inbox

Thank you!

Around the globe