BURSTNothing is more fundamentally human than identity. How we think about ourselves shapes our experiences and how we navigate the world.
Yet, the idea of identity existing as a unique quality anchored to a specific individual is undergoing change. Artificial intelligence (AI), virtual reality, the metaverse, and other digital tools make it possible to experience more dynamic, contextual, and exchangeable identities that let people walk in the shoes of others and tap into their feelings, perceptions, and knowledge.
“Historically, identity attributes have been missing from databases. However, with today’s technologies we are able to build a representation of a person’s identity,” said Hossein Rahnama, a visiting associate professor at the Massachusetts Institute of Technology (MIT) Media Lab, and associate professor at Canada’s Toronto Metropolitan University.
These digital identities can be rented, traded, and exchanged. The technology can provide a window into another person’s thinking, allow users to walk in the shoes of a historical figure, and even let them experience the emotions and feelings of others.
ID, Ego, and AI
Today, identity characteristics typically exist in vague notions or in structured databases that include fixed attributes like name, gender, and age, as well as things like interests, hobbies, and lifestyle preferences. The complexity and fluidity of human behavior makes it largely impossible to replicate a person’s true identity and behaviors.
That is changing. At the MIT Media Lab, Rahnama is pushing the boundaries of identity within the digital realm. He and his research team have developed a framework called Perspective Aware AI, which encapsulates behavioral patterns and personality traits through a new type of data structure called Human-AI Chronicles. These digital representations—think of them as a digital twin—behave and interact in human-like ways.
These swappable identities are more than the basic avatars that appear in virtual reality and gaming. They generate interactive written chats, images, and speech. “It transforms the individual’s identity and wisdom into a portable framework. It expands the idea of collective intelligence,” Rahnama said. “This makes it possible to build networks where people can access each other’s knowledge and ‘being’ in intimate and accessible ways.”
This knowledge graph feeds data about an individual’s personality characteristics to the AI framework. This data comes from voice and video recordings, written materials, sensors on smart watches, and more. Rather than storing data in a structured database, the knowledge graph creates a web of dynamic and interconnected attributes based on causalities. These subgraphs—mathematical representations of various personality components—allow AI to evolve beyond stored facts and gain robust inferencing and reasoning capabilities.
The result is an AI-generated representation of a person that can interact via a chatbot, on a 2D screen, or in a 3D virtual reality environment. This system simulates an individual’s thought process and behavior through an ultrarealistic representation. What’s more, “The knowledge graph can evolve over time and become richer and more detailed,” Rahnama said.
Virtually There
Exchangeable digital identities could lead to new types of apps and interactions. Digital twins—hyper-realistic versions of real people—could become coaches, mentors, trainers, or companions. In a workplace, this could allow a CEO or knowledge guru to interact with employees in new ways.
It would also be possible to buy time with a virtual celebrity or walk beside an historical figure like Rosa Parks or Neil Armstrong. Parents and grandparents could share their wisdom and provide advice, even after they die. “The technology allows people to share knowledge and improve decision-making, but it also establishes stronger connections and promotes empathy,” Rahnama said.
Jeremy Bailenson, Thomas More Storke Professor of Communication at Stanford University, has spent years studying identities within virtual reality environments. In 2006, his Virtual Human Interactions Lab at Stanford coined the term Proteus Effect to describe a psychological phenomenon in which a person’s online and offline behavior can be shaped by a digital avatar or digital representation.
In practical terms, this means that a person can gain confidence and become more assertive, empathetic, and compassionate by spending time with an avatar or digital identity. “What avatar one dons during gaming, socializing, and work in online platforms has consequences,” Bailenson said. For example, in one study he conducted, color-blind participants were found to be more likely to aid others.
Mel Slater, a distinguished investigator at Spain’s University of Barcelona, has found that people who spend time with avatars and digital identities in virtual reality environments begin to perceive the character as part of their own body. In one study, Slater discovered that white participants who embodied Black avatars in VR displayed long-term reductions in implicit racial bias.
The opportunity to expand thinking and consciousness in new ways could be nothing short of revolutionary. “Research shows that an immersive experience, where you actually feel as though you are in the body of someone else, or feel as if you’ve taken on a new ability, can especially impact your thoughts and behaviors in the real world,” Bailenson said.
Identity Matters
While digital identities could profoundly reshape the way humans learn, interact, and behave, they also raise a slew of ethical questions. This includes ownership rights for digital personas that outlive their creators, the rights and responsibilities associated with a digital self, and the legal consequences for information or behavior that leads to injury or death. Still another critical issue is how to protect personal data from fraud, manipulation, or deep fakes.
Developing protocols that allow individuals to control their data and how it is used is essential, Rahnama said. Chronicles, for instance, leverages the Open Algorithms Project (OPAL). It stores data centrally but grants algorithmic access based on defined policies. “Individuals can choose the granularity of data sharing and set varying trust levels depending on whether the requester is familiar or a stranger,” Rahnama explained.
Although it’s impossible to predict exactly how digital identities will evolve, it’s clear that they will reshape the way we interact with computers, AI, and other humans. The hope is that they will enable more personalized, empathetic, and participatory interactions, and open doors to new forms of entertainment, learning, and knowledge. Said Rahnama, “The goal is to allow people to interact in new and richer ways—and understand each other better.”
Samuel Greengard is an author and journalist based in West Linn, OR, USA.