According to a study by Proofpoint, unveiled at the Web Summit 2026 in Lisbon, the number of detected phishing attempts (fraudulent messages aiming to obtain credentials or data) has been multiplied by 4.5 in Europe in one year, compared to 1.5 globally. The report also notes an intensification of state-linked espionage operations and a growing targeting of cloud and AI infrastructures.
Researchers from the company, which presents itself as a global leader in “human- and AI-agent-centric” cybersecurity, observed signs of successful compromise in 64% of European organizations targeted by attacks against their cloud environments. These attacks rely notably on phishing and session hijacking, which allow access to specific users and accounts. According to the study, their effectiveness illustrates the limits of more general brute-force attacks.
Users Clicking More Often
In Europe, users clicked on 8.3 malicious messages per 100 threat messages delivered, compared to 2 globally. This finding, according to the report, underscores the still decisive role of the human factor in cybersecurity.
Proofpoint’s Chief Security Officer (CSO), Ryan Kalember, stated: “The cyber threat landscape in Europe is evolving rapidly. Phishing attacks and those targeting digital identities remain among the most significant risks, but attackers are also expanding their scope to new AI infrastructures.” He adds: “What is particularly concerning is the increase in the volume and sophistication of these attacks. AI enables cybercriminals to operate at an unprecedented scale.”
Cloud and AI, New Attack Grounds
Proofpoint indicates it has identified more than 188,000 applications integrating AI features and connected via authentication mechanisms and OAuth infrastructures, a protocol allowing an application to access services without directly sharing the user’s password. Furthermore, the proportion of European sensors affected by attacks targeting open-source AI technologies was multiplied by 4.6 between two consecutive 90-day periods.
Kalember believes organizations must now “protect themselves against current threats while anticipating new risks linked to work environments integrating autonomous AI agents.” Humans and AI, he says, “are increasingly accessing the same systems and the same sensitive data.”
The report also describes threats specific to certain sectors. In hospitality, tourism, and leisure, malware represents 21% of detected threats, nearly double the global rate. In European healthcare, callback phishing increased by 257% in one year, and malware detections by 84%. Transport and logistics companies, for their part, can be exposed to attacks facilitating cargo theft, while sophisticated campaigns target investors and consumers, notably by impersonating the French national health insurance.
The study’s authors conclude with a more general evolution: attackers are increasingly targeting the interactions between people, data, and AI, as organizations accelerate the adoption of artificial intelligence.
