Researchers seek to mimic digital identities by analyzing email, online interactions
Your digital paper trail could tell enough about your thoughts, interests, and personal identity that an artificial intelligence chatbot could emulate it, new research seeks to prove.
Research being done at the MIT Media Lab is working on “swappable identities” for AI bots, based on data taken from a person’s digital identity, as detailed by VentureBeat. Personal information is culled from emails, transcribed videos and any other published statements, allowing the system to give expert advice based on human opinions.
MIT Media Lab sees swappable identities being advantageous in a number of ways, providing examples of asking for a digital Donald Trump’s opinion on a matter that might affect a company’s revenue stream, or entrepreneurs seeking the advice of Tesla founder Elon Musk.
Another example could be company managers, who could be made available around the clock for team members to ask for guidance. A swappable identity could reduce the need for time consuming meetings, allowing employees to be more productive.
Of course, such technology also raises privacy and security concerns, not the least of which could be a new form of digital identity theft. Imagine a swappable identity chatbot that could impersonate you via email or text message as a means to nefarious ends.