OpenAI launches Lockdown Mode and Elevated Risk warnings to protect ChatGPT against prompt-injection attacks and reduce data-exfiltration risks.
These 4 critical AI vulnerabilities are being exploited faster than defenders can respond ...