The message arrives as a grey panel on an iPhone lock screen and reads like a fire alarm: Apple has detected activity consistent with a targeted attack, and the phone may be compromised. Most recipients dismiss it. This weekend, a number of them were members of the Bundestag.
Germany’s federal cyber security agency, known as BSI, issued an advisory over the weekend saying that members of parliament and several public figures had received threat notifications from Apple warning that their devices may have been infiltrated by government-backed spyware. Berlin said it was taking the matter seriously. Apple’s notifications in this round went to users in 99 countries, according to the advisory and the company’s own account of the campaign.
Its role is to assess and coordinate rather than to investigate a criminal case, and its warning was carefully worded. The agency did not say which government might be responsible, did not confirm that any device had in fact been breached, and did not describe the mechanism of infection. Apple, for its part, says it does not attribute attacks to specific governments and does not confirm individual compromises, on the grounds that doing so would help the attackers learn what is being detected.
The gap between a warning and a finding is where most of these cases live. Threat notifications are triggered by signals Apple will not fully disclose: anomalous processes, known exploit artifacts, patterns of outbound traffic that resemble a targeted intrusion. They are, by design, probabilistic. A lawmaker who receives one cannot tell from the screen whether the phone in their pocket has been running someone else’s commands for a month or whether a researcher’s heuristic fired on a false positive.
The tools involved in cases like this are sold commercially and priced for states. The most documented is Pegasus, built by the Israeli firm NSO Group, which infected phones without any action by the user and turned them into listening devices that could read messages, follow location and open the microphone. Similar products from other vendors have been found on devices in Europe, Asia and Latin America. Germany has spent years arguing about whether its own police forces should be permitted to use them, a debate that becomes awkward when the targets are legislators who vote on the laws governing surveillance.
The technical vectors are narrow but durable. Modern iPhones run a hardened messaging pipeline and a lockdown mode that strips out the features most exploits rely on, and Apple has closed a long list of remotely triggerable flaws. Each closure moves attackers toward the next unpatched path, or toward tricking a user into installing a malicious configuration profile or a fake app. A targeted campaign needs only one of those doors to stay open for a few weeks.
Attribution usually comes from outside the phone. Researchers at Citizen Lab, Google’s threat analysis group and a handful of nonprofit labs match code fragments, infrastructure and operator mistakes to specific clients, and publish findings that force governments to respond. That work takes months. In the interval, the practical advice for a member of parliament is to replace the handset, avoid carrying it into meetings where classified material is discussed, and assume that anything that was on the device is now known to someone else.
The political dimension is not incidental. Germany’s parliament has been penetrated before: its internal network was breached in 2015 in an intrusion that federal prosecutors later attributed to Russian intelligence, and the cleanup took years. A lawmaker’s phone is valuable less for its messages than for its position in a web of contacts, calendars and group chats that map a committee’s work. Spyware on a single device can reveal who is drafting what, which witnesses are cooperating and when a minister is about to change position.
For the agency, the advisory is also a message to the institutions around the affected individuals. Government departments, party offices and parliamentary administrators have to decide whether to harden devices, hand out replacements or accept the risk. Those decisions cost money and create visible inconvenience, and they are usually deferred until an intrusion is confirmed. By then the interesting traffic has already left the building.
Harder still is the question of what a notification obliges anyone to do. A warning that cannot be verified leaves the recipient with two bad options: ignore it and hope, or rebuild a working life around a new handset, new numbers and lost message history. Most people, including most politicians, do the first until someone with a badge tells them otherwise.
Apple has shipped these notifications since 2021, and their volume has risen as the mercenary spyware market has grown. The company has been criticized for saying too little about what triggers one; it argues that saying more would let operators tune their tools to stay under the threshold. Both positions are reasonable, and neither helps the person staring at a grey box on a lock screen at seven in the morning, deciding whether to call the office.
