Artificial Intelligence Safety and Cybersecurity: a Timeline of AI Failures
This work addresses AI safety concerns for researchers and policymakers by highlighting the escalating risks, though it is incremental as it builds on existing cybersecurity concepts.
The paper analyzes reported failures of AI systems and projects that both the frequency and severity of future AI failures will increase, with narrow AI failures being moderate like cybersecurity issues, but general AI failures potentially causing catastrophic events without recovery.
In this work, we present and analyze reported failures of artificially intelligent systems and extrapolate our analysis to future AIs. We suggest that both the frequency and the seriousness of future AI failures will steadily increase. AI Safety can be improved based on ideas developed by cybersecurity experts. For narrow AIs safety failures are at the same, moderate, level of criticality as in cybersecurity, however for general AI, failures have a fundamentally different impact. A single failure of a superintelligent system may cause a catastrophic event without a chance for recovery. The goal of cybersecurity is to reduce the number of successful attacks on the system; the goal of AI Safety is to make sure zero attacks succeed in bypassing the safety mechanisms. Unfortunately, such a level of performance is unachievable. Every security system will eventually fail; there is no such thing as a 100% secure system.