Anthropic Security Chief Warns Cyberattacks Will Worsen

Cybersecurity leaders face a growing threat from increasingly capable artificial intelligence tools, Anthropic Deputy Chief Information Security Officer Jason D. Clinton warned on October 22, 2026. His assessment follows a series of data breaches and disruptive attacks affecting organizations across Europe, as attackers exploit software weaknesses and gaps in security defenses.

In a post on X, Clinton cautioned that the situation could deteriorate before improved defensive tools become widely available. He urged companies to prioritize fundamental security measures and secure support from their boards rather than focusing exclusively on AI-related initiatives.

"As bad as things have gotten in the past 4 days, they are going to get a lot worse before they get better," said Jason D. Clinton, Anthropic's Deputy Chief Information Security Officer.

Clinton called on security leaders to use the current crisis to obtain additional organizational support. His warning reflects concerns that businesses may struggle to identify, patch, and defend against vulnerabilities quickly enough as attack methods become more automated.

"Security leaders, never let a crisis go to waste: please go to your boards and ask for help right now. Not for AI, for the basics." wrote Jason D. Clinton, Anthropic's Deputy Chief Information Security Officer.

Clinton also reiterated a warning he had issued in August about an approaching surge in software vulnerabilities. He had cautioned that critical infrastructure operators might not have sufficient time to patch serious weaknesses as the number of exploitable flaws increased.

He subsequently highlighted the concept of "vulnzero," referring to the goal of eliminating software vulnerabilities. Clinton expects advanced AI security tools to become less expensive, potentially allowing developers to identify and correct weaknesses before software enters production.

Until those tools become common in everyday development, however, attackers can use existing AI capabilities to identify weaknesses in older software, automate security flaw discovery, and exploit known vulnerabilities before defenders deploy patches.

The warning comes amid a series of cyber incidents reported across Europe in recent weeks. According to cybersecurity company Check Point, global attack attempts increased 48% in September 2026 compared with the same month a year earlier.

In Switzerland, attacks affected the federal pension fund Publica, the Swiss Broadcasting Corporation (SRG), online magazine Republik, and space company Beyond Gravity. The incidents exposed weaknesses in software suppliers and third-party services, demonstrating how attackers can reach organizations through their technology partners.

Publica's software vendor, PK Softech, was targeted with malware, potentially exposing salary information, retirement savings, account numbers, addresses, social security numbers, and spousal details. The City of Zurich's pension fund was also affected.

Republik disclosed that attackers exploited a vulnerability in the Metabase analytics tool. A patch had been available since August 6, 2026, but the magazine did not install it in time. Names and email addresses of current and former readers were accessed, along with additional personal details for some individuals.

Other incidents demonstrated the potential scale of compromised information. Denmark's central population register suffered the theft of approximately 8.8 million records after attackers exploited access through a private IT company. In Britain, retailer ASOS reported a breach involving customer information, while a separate attack affecting Discord's third-party security provider exposed account-related data.

Cybersecurity experts have also pointed to expanding digital services and increasingly complex technology dependencies as factors increasing exposure to attacks. Supply chain breaches can allow attackers to compromise a target indirectly by exploiting weaknesses in a software vendor or another connected service.

Expanded incident-reporting requirements, including rules under the European Union's NIS2 and DORA frameworks, may also contribute to the growing volume of public reports. Organizations that notify authorities about incidents may disclose them publicly for transparency, increasing the number of attacks reported in the media.

Clinton urged the technology industry to prepare for an extended period of heightened risk while stronger defensive capabilities develop. In his October 22 post, he called on security leaders to maintain their focus for approximately 18 months as AI-driven attack capabilities continue to evolve.