The danger of these technologies lies in their ability to automate multiple stages of an attack, from gathering information about organizations to drafting phishing messages and developing malicious software, which may allow attackers to target a larger number of victims with fewer resources.
According to a report issued by the Google Threat Analysis Group in November 2025, threat actors have used artificial intelligence in tasks including reconnaissance, phishing, malware development, and data extraction attempts, with prototypes of software capable of using language models to modify their behavior during operation being observed.
Fraudulent messages have also become more sophisticated thanks to the ability of smart models to draft linguistically sound texts tailored to the nature of the victim, making them difficult to detect based on the spelling errors or poor phrasing that were among the most prominent signs of traditional phishing.
However, the proliferation of these tools does not mean that hacking complex systems has become accessible to everyone; exploiting vulnerabilities, bypassing protection measures, and maintaining access to networks still require, in many cases, technical expertise and careful planning.
As an example of the development of these capabilities, Anthropic announced in November 2025 that it had detected a cyber-espionage campaign it said was carried out with the help of the "Claude Code" tool, targeting about 30 organizations around the world, with confirmed breaches in a limited number of them. The company explained that the tool helped automate parts of the process, while humans retained a role in selecting targets and directing the attack.
The repercussions of these attacks are not limited to data theft, but include financial fraud, customer identity theft, and extortion, as well as damage to the reputation of institutions and the trust of their clients, especially when the vulnerabilities are in the systems of third-party service providers.
In the face of these risks, the importance of updating software, closing known vulnerabilities, enabling multi-factor authentication, restricting access privileges, and monitoring unusual activities stands out, in addition to maintaining secure backups and being prepared for rapid incident response.
Conversely, organizations can leverage artificial intelligence to analyze security logs, triage alerts, detect suspicious behavior, and accelerate threat response, provided that human oversight and appropriate controls are in place.
Thus, artificial intelligence does not make hacking organizations easy or guaranteed, but it lowers the barriers for some attackers and gives them the ability to carry out faster and broader operations, making strengthening cyber defenses and early detection of attacks an increasing necessity.
