Archive — US — techcrunch
Signal President Meredith Whittaker calls out agentic AI as having ‘profound’security and privacy issues
At SXSW, Signal President Meredith Whittaker highlighted the significant privacy and security risks associated with agentic AI, warning that these AI systems need extensive access to user data, including credit cards and messaging apps, potentially breaking privacy boundaries and increasing vulnerability to surveillance.
Key points
- Meredith Whittaker warns that agentic AI poses privacy and security risks due to its access needs.
- AI agents can perform various tasks for users but require extensive data and system access.
- Accessing sensitive information, like calendars and credit card details, raises security concerns.
- Whittaker notes that powerful AI models likely require cloud processing, heightening privacy risks.
- Integrating AI agents into apps like Signal could undermine message privacy.
Why it matters
- Whittaker's concerns reflect broader debates about AI ethics and data privacy amidst increasing AI adoption.
- The growing trend of using AI in daily activities relates to concerns over data security seen in previous technology integrations.
- Historical precedents of surveillance and data misuse highlight the potential risks Whittaker describes.
Whittaker's warnings about agentic AI underscore ongoing security and privacy challenges, emphasizing the need for careful consideration and regulation as AI technology becomes more integrated into our lives.