Open AI models are catching up fast
Public AI models are getting better at cyber tasks at a surprising speed. Research from the British AI Security Institute shows that these models now trail the best paid tools by only a few months. Early last year the gap was much wider.
The cost to run these powerful public tools is also dropping quickly. This makes high-level AI capabilities available to far more people than before. It means that advanced tools are no longer limited to only a few large companies.
Safety features on these public models are still not very effective. This leaves those who defend digital systems with very little time to prepare. The rapid pace of this technology is creating new risks for online security.
Comments (0)
No comments yet. Be the first!
More AI news
NewsChina Launches New Global AI Group
China is setting up its own AI governance network to challenge Western influence.
NewsGoogle Cloud Ditches Vector Databases for Continuous Memory
Google Cloud has launched a new memory agent that manages data without using traditional vector databases or embeddings.
NewsThe Navy is going all-in on AI
The US Navy just launched a new plan to put AI tools directly onto its warships.