AI Risks Making Software Too Secure, Threatening Law Enforcement
TL;DR. New analysis suggests artificial intelligence could paradoxically increase software security to a point that undermines law enforcement's ability to access data. - AI-driven security enhancements may render traditional surveillance and hacking methods ineffective for intelligence agencies. - The 'Going Dark' problem, previously driven by encryption, re-emerges with AI's potential to harden software against intrusion. - This development raises concerns for national security and also for the broader computer security and privacy landscape.
- AI is poised to significantly increase software security.
- This heightened security could make intelligence and law enforcement agencies 'go dark'.
- The inability to access digital data poses a challenge for national security.
- The development also impacts the broader computer security and privacy discourse.
Sources
- Everything is about to "go dark" — blog.cryptographyengineering.com