The rapidly increasing capabilities of AI now seem to be a new challenge for cyber security. AI company Anthropic has decided not to make public its new model “Cloud Mythos”. According to the company, this model is capable of identifying and exploiting hidden vulnerabilities in decades-old software systems. Anthropic claims that other AI labs are still 6 to 18 months behind in achieving this level of capability. This has increased concerns that large parts of the Internet could become more vulnerable. Whether it is a streaming platform, online banking or a search engine, the trend of creating software without special training has also increased, due to which such applications are increasingly being created which do not undergo security testing. As AI becomes more advanced, it will become easier to find and misuse them. Till now the Internet has been relatively safe because the software was complex but open. This balance is changing amid increasing reliance on source software. For example, systems like FFMPEG and security-related OpenBSD are running on limited resources. According to Anthropic, experts have discovered 27-year-old vulnerabilities in FFMPEG. Hackers can target hospitals, networks and critical infrastructure. In this situation, experts insist that cyber security should not be an option but should be made the default. Companies will have to include security measures in their software in advance, so that the risk can be controlled in this new era of AI. Make safety the default. Companies that include open source code in their products should hire necessary workers for maintenance. Companies that make tools like Mythos should do this. Hand these tools to the workers. These safety measures need to be a default instead of a premium feature for millions of new creators.
Source link