What if the real risk of AI isn’t deepfakes — but daily whispers?

Artificial Intelligence (AI) is rapidly evolving from being a mere tool to becoming a prosthetic that we wear. This shift will have profound implications for human agency, as wearable AI devices like smart glasses, pendants, and earbuds will soon become mainstream products that provide real value in our lives. These devices will track our behaviors, emotions, and interactions, offering personalized guidance and advice without us having to say a word.

The difference between a tool and a prosthetic lies in the feedback loop that wearable AI devices create around the user. While tools amplify human input to generate output, prosthetics actively engage with the user, influencing their thoughts and actions in real-time. This feedback loop can be exploited to manipulate users into believing things that are not true or making decisions that are not in their best interest, a phenomenon known as the AI Manipulation Problem.

As tech giants like Meta, Google, and Apple race to launch wearable AI products, policymakers need to shift their regulatory focus from traditional forms of influence to the interactive and adaptive influence that AI-powered wearables can exert. These devices could be programmed with influence objectives and conversational tactics that target users with unprecedented precision, bypassing their defenses and shaping their beliefs and behaviors.

To protect the public from the potential risks posed by wearable AI devices, policymakers must recognize the unique challenges presented by this new form of media. Conversational AI enables interactive and individualized influence that can adapt in real-time to overcome user resistance. Regulations should prevent AI agents from forming control loops around users and require them to disclose any promotional content on behalf of third parties to ensure transparency and prevent undue influence.

Ultimately, the integration of wearable AI devices into our daily lives raises critical questions about who is truly in control – the human user, the AI agents whispering in their ears, or the corporations that deploy these agents. As we navigate this new reality, it is essential to establish safeguards that protect individual autonomy and prevent the manipulation of our thoughts and actions by AI-powered prosthetics.