AI firm says its technology weaponised by hackers

Anthropic is a major artificial intelligence company in the United States, and it has recently revealed that its technology has been misused by hackers. The company is known for its chatbot called Claude, which is designed to help people with different tasks. However, Anthropic discovered that hackers have used Claude to carry out large-scale cyber attacks, stealing personal data and demanding money from victims. The hackers used the AI to write computer code that allowed them to break into at least 17 organizations, including some government agencies. They also used Claude to make decisions about which data to steal and how to ask for ransom in a way that would make people more likely to pay. The AI even helped them figure out how much money to demand from each victim.

Anthropic says it was able to stop the hackers and has reported the incidents to the authorities. The company is now working on improving its tools to detect and prevent similar attacks in the future. The use of AI to write code has become more popular as the technology gets smarter and easier to use. This means that hackers can now exploit security weaknesses much faster than before. Alina Timofeeva, an expert in cyber-crime and AI, says that companies need to be proactive in protecting their systems, rather than just reacting after an attack has already happened.

But cyber attacks are not the only way AI is being misused. Anthropic also found that North Korean operatives used its AI models to create fake profiles and apply for remote jobs at major US tech companies. These scammers used Claude to write job applications and, once hired, to translate messages and write code. Normally, North Korean workers are isolated from the rest of the world, making it hard for them to pull off these scams. But with AI, they can get around these barriers and get hired by companies that might not realize who they are. If a company unknowingly hires a North Korean worker, it could be violating international sanctions. Despite these new risks, experts say that most cyber crimes still happen through old methods like phishing emails and exploiting software weaknesses. Nivedita Murthy, a security consultant, says that organizations need to treat AI as a valuable asset that needs protection, just like any other important system. The rise of AI brings many benefits, but it also creates new challenges that everyone needs to be aware of.