Cyberattack reports rise, but successful attacks may not

Reports of cyberattacks on Swiss companies have increased, but that does not necessarily mean successful attacks have become more frequent, SRF reports. Several pension funds may have been affected by a data theft after an attack on a software provider whose programs they use to manage clients. Earlier attacks on Ricardo and SRG had also been reported.

More incidents are being reported

Cyberattacks happen constantly, SRF digital editor Reto Widmer said, but most are unsuccessful and go unnoticed. The number of reports rose from seven in the month before Switzerland introduced a reporting obligation to more than 20 in the month after. There were 63 reports in September.

The increase may partly reflect companies being more open about incidents. In the past, businesses may have been more reluctant to disclose an attack, Widmer said. Companies now tend to inform the public sooner, and well-known firms have also been targeted, making incidents more visible.

Since April 1, 2025, operators of critical infrastructure have had to report certain cyberattacks, including incidents involving stolen data. The Federal Office for Cybersecurity said it received 164 reports of attacks on critical infrastructure between the start of April and the end of September, with the financial sector accounting for the most. Operators that fail to report can face fines of up to 100,000 Swiss francs.

AI is not needed for many attacks

Widmer cautioned against assuming that artificial intelligence is behind every cyberattack. AI can help identify software vulnerabilities by analysing large amounts of code, but attackers often rely on tools that already exist.

Widely used chatbots such as ChatGPT have safeguards intended to prevent them from assisting with cyberattacks. So-called malicious AI models are designed for harmful or criminal use, but operating them requires time and expertise. They can also be rented, although the computing power costs money.

Many companies still have weak cyber hygiene and have not fixed vulnerabilities that have been known for years. Attackers can exploit these flaws without AI, or rent ready-made phishing kits for a few hundred dollars a month. Strong cybersecurity can make attacks difficult even for AI agents, Widmer said.