Personal AI twins are being used to speak and decide in your style

In short: More people and products are experimenting with “AI twins,” which are AI versions of a person that can talk like them and use their past words to respond.

What's going on

Some apps and public demos now offer personal “AI twins,” sometimes called digital twins or AI clones. The idea is simple, if you are busy, someone can talk to an AI stand-in that tries to answer the way you would.

These systems usually start with a general AI text model, then add your own material on top. That personal material can include your emails, notes, articles, or transcripts. A common setup is a “retrieval” step, which is like a built-in search tool that pulls up your own past words before the AI replies.

One consumer example is COPYMIND, which markets itself as “your personal AI twin.” It is positioned as a private self-reflection tool, where conversations help the model learn your patterns, goals, and preferences. The company also highlights always-on availability and privacy claims.

On the public figure side, LinkedIn co-founder Reid Hoffman has shown a digital twin called “Reid AI.” He says it has joined him in judging a Stanford business plan competition, appeared in TV interviews, and helped deliver a commencement speech in 10 languages. Hoffman also noted limits, including occasional awkward responses and less nuanced advice.

What to watch

Two issues will likely shape what happens next. First is trust and transparency, meaning people should know when they are talking to an AI, not the real person. Second is control and safety, since the same tools that help someone “clone themselves” could also be used for impersonation if consent and identity checks are weak.

Source: NYTimes