The findings highlight the urgent need for comprehensive security measures in AI systems to prevent data leaks and protect user privacy. The post TikTok researchers warn that MCP tools can quietly turn AI agents against their users appeared first on Crypto Briefing.