“To reap the advantages of AI, customers should have confidence that the AI will behave as designed, and outcomes are secure and safe,” the CSA stated within the information. “Nevertheless, along with security dangers, AI programs may be susceptible to adversarial assaults, the place malicious actors deliberately manipulate or deceive the AI system.”
The rules don’t give attention to AI security or broader points generally related to AI, corresponding to equity, transparency, or inclusion, nor do they sort out cybersecurity dangers launched by AI programs.
Whereas some really useful actions might overlap with these areas, the rules additionally don’t particularly tackle the misuse of AI in cyberattacks, corresponding to AI-enabled malware, or threats like misinformation, disinformation, and deepfakes, the CSA stated.
