Hello my dears,
after studying the report, I wonder:
I'd love to hear your thoughts in the comments.
Elevators are increasingly being controlled digitally, making them a potential target for cybercriminal attacks. According to a Tüv report, many systems are not sufficiently protected against this.
Washington wanted to use the Anthropic AI "Mythos" for autonomous weapons and surveillance, but the company refused and was classified as a security risk. A meeting has now taken place.
The US government and the head of the AI company Anthropic, Dario Amodei, have met for talks for the first time since a dispute at the beginning of the year. According to the White House, the meeting on Friday focused on possible cooperation and joint approaches to overcoming the challenges posed by Anthropic's AI models.
Anthropic described the meeting as "productive". They discussed how to work together on key issues such as cyber security, the leadership role of the USA in the race for Künstliche Intelligenz (AI) and AI security. The background to this are growing fears that Anthropic's latest AI model, "Mythos", could significantly increase cyber attacks.
"Mythos" detects software vulnerabilities
Anthropic's new AI has alarmed cyber security experts worldwide. It enables non-experts to launch complex hacker attacks. In initial tests, "Mythos" discovered security vulnerabilities that had remained undetected for years despite numerous tests by humans.
The banking sector is considered particularly vulnerable to attacks by AI due to its outdated technology systems. Government representatives from the USA, Kanada and Großbritannien have met with leading bank representatives to discuss the threats posed by "Mythos". The Europäische Zentralbank (ECB) wants to discuss possible consequences of this development with leading bank managers, according to insiders.
The Anthropic "Mythos" model announced on April 7 will initially be made available to a few selected companies. As part of the "Project Glasswing" initiative, companies will be allowed to use the model to search for weaknesses in cyber security.
Anthropics neues Modell: US-Regierung trifft CEO wegen Hacker-KI
