In a recent interview with Bloomberg, Meredith Whittaker, President of Signal, addressed the growing concerns surrounding the privacy implications of AI chatbots like ChatGPT and Claude. She emphasized that these technologies should not be perceived as friends or conscious entities. “These are not your friends. These are not conscious beings. These are not sentient interlocutors,” she stated, underscoring the necessity of maintaining a critical viewpoint towards AI interactions.
Whittaker shared her cautious approach to using AI tools, admitting that while she occasionally utilizes them for minor tasks such as formatting documents, she refrains from asking them questions. “I’m very serious about my thinking and writing,” she explained, expressing her belief that the AI’s responses could hinder the authentic process of developing ideas.
The discussion progressed to Microsoft CEO Mustafa Suleyman’s prediction that users might rely on Microsoft Copilot for their Christmas shopping this year. Whittaker expressed skepticism about such a scenario where an AI would potentially monitor family conversations to gauge shopping preferences. She raised significant concerns regarding data security, noting that this would require broad access to personal information, including credit card details, browser history, messaging apps, home addresses, and calendars.
Describing this level of access, Whittaker remarked, “What you’ve just described is a system with very pervasive access across multiple applications and services.” She argued that within the context of Signal, this type of access would amount to creating a backdoor, fundamentally compromising user privacy.
Whittaker’s comments reflect a growing apprehension among experts regarding the implications of integrating AI tools into everyday life. As discussion around AI continues to evolve, her perspective serves as a reminder of the importance of prioritizing privacy and conscious engagement with technology.
