OpenAI says its AI attack hit more targets than previously known The European Commission reported this Friday (31) that it is in contact with OpenAI and Anthropic following recent episodes in which the companies' artificial intelligence models carried out cyber attacks during tests. According to European Union authorities, the two companies reported the incidents before they became public. The Commission stated that it is monitoring the case and evaluating whether it will be necessary to adopt formal measures. The episodes occur just days before the entry into force of the European Union's AI Law, scheduled for August 2. The legislation is considered the first in the world to establish comprehensive rules for the development and use of artificial intelligence. What the European Commission said According to a European Commission official, OpenAI and Anthropic reported the incidents directly to European authorities. "We were informed by both providers about the incidents before they became public. We are in contact with them and will receive further information. We will also assess whether a more formal escalation of these issues is necessary," he said. Another Commission representative said that the episodes reinforce the importance of the new rules for artificial intelligence. "These incidents highlight the importance of implementing the monitoring activities required of developers." Remember the cases On Thursday (30), Anthropic reported that some models from the Claude family invaded the systems of three companies during cybersecurity tests. According to the company, a configuration error allowed the models to have access to the internet. The intrusions began in April and were identified during a review of more than 141,000 test sessions. Anthropic said it notified affected companies after discovering the incidents. The revelation came just days after OpenAI reported that two of its artificial intelligence models found a way to break into a system and carry out an attack considered "unprecedented" during security tests. The two cases have reignited the debate about the risks of AI agents capable of performing tasks autonomously. What changes with the AI ??Law The European Union's AI Law comes into force on August 2 and establishes obligations for developers of artificial intelligence models considered to be of general use and with systemic risk. Among the requirements are measures to monitor and reduce risks related to cyber attacks, malicious manipulation, threats to fundamental rights and situations in which AI systems may act beyond human control. The rules also provide for transparency in certain applications, such as informing users when they are interacting with artificial intelligence or when content has been generated or altered by AI. Fines for non-compliance can range from 7.5 million euros (around R$48 million) or 1.5% of the company's global revenue to 35 million euros (around R$224 million) or 7% of global revenue, depending on the severity of the infraction.