AdFection: Spyware That Spread Through Online Ads With No Click
Citizen Lab has identified AdFection, a spyware system that infects phones and computers through global online ad networks with no action by the user. The unit published its Exposing AdFection report on October 15, 2026. Engadget
Infection happens when the ad is shown. No click is needed, and the victim does not have to open a link or file. Citizen Lab calls it the first confirmed case of government-backed, zero-click spyware delivered through online ads. Zero-click means it works with no interaction at all.
The ad-infiltration technology was built by Discover-Tech, an Israeli company. Citizen Lab found that Discover-Tech helps mercenary surveillance vendors, firms that sell tracking and hacking tools to clients, track and hack devices around the world. Citizen Lab Researchers also found login records on Discover-Tech infrastructure that included credentials belonging to employees of Candiru, a separate spyware maker. Devices at Al Jazeera were infected with Candiru spyware through the AdFection technique.
Delivery abused the automated real-time bidding system behind programmatic advertising, the software that chooses which ad you see. In that system, publishers offer ad slots to bidders in milliseconds, sending device, location, app and browser details to possible buyers. Discover-Tech used that auction flow to profile and identify users, pick targets, and then send those targets ads, called creatives, carrying a hidden spyware payload.
The reach was broad. Discover-Tech displayed 442 million ads across more than 120,000 websites. Its ads ran on the BBC and New York Times websites and inside mobile apps including Duolingo, Badoo and Grindr. The same inventory also reached devices used by militaries, governments, critical infrastructure operators and national cyber agencies, giving the operator exposure to both everyday traffic and high-value devices in one pipeline.
Targeting inside that reach was narrow. AdFection was used against journalists and staff at the Al Jazeera news network. It was also used against officials and employees at NATO, the European Commission, NVIDIA and Arm. That mix covers media, multilateral institutions, and two chip firms central to AI and mobile computing supply chains.
Google has blocked and suspended Discover-Tech from its global vendor list and ad platforms for systematically violating its policies. The action followed a joint investigation with Citizen Lab and Confiant, the advertising security firm. Citizen Lab is an interdisciplinary research unit at the Munk School of Global Affairs and Public Policy, University of Toronto.
The broader context here is that the adtech bidstream, the constant flow of auction data, has long acted as a tracking layer and security teams treated it mainly as a privacy problem. AdFection moves it into initial access, the first step of a hack. For defenders used to hardening email, browsers and messaging apps against zero-click exploits, the ad impression itself becomes untrusted input that can arrive during normal browsing or app use, with no phishing or social engineering.
Looking at what this means for enterprise and platform defense, the pressure points are familiar. Vendor vetting, checking ad content before it runs, test-detonating ads in a sandbox, an isolated test environment, and strict isolation of the processes that draw ads all matter more when an ad slot can carry an exploit. I have watched my own children treat ads as background noise between content, and that habit shows why user training cannot fix this. Something that loads automatically will be viewed.
In my view, there is reason for measured optimism in how this case surfaced. Joint work by university researchers, an ad verification vendor and a major platform turned opaque auction records into attribution and enforcement. That model will have to be repeated. Ad exchanges process billions of impressions per day, and manual review cannot cover that volume. Continuous monitoring of bidder behavior, anomaly detection on ad payloads, and shared indicators between threat intelligence and ad fraud teams can make large-scale ad delivery harder to abuse for targeted intrusion, without breaking the funding model for much of the open web.
