The line between humans and machines is blurring—faster than we ever imagined. In just two hours, artificial intelligence can create a digital replica of a person that mimics their personality with 85% accuracy. What sounds like science fiction is now reality, thanks to a groundbreaking study by researchers at Stanford University and Google. They have developed “simulation agents”—AI models capable of replicating human behaviour with astonishing precision. But what does this mean for society? And where do we draw the line between innovation and ethical uncertainty?
The Birth of a Digital Twin
The researchers conducted two-hour interviews with 1,052 participants, delving into their life stories, values, and social views. These conversations fed into a generative AI system, which then created a personalised simulation model for each person. The participants and their AI counterparts were then put through personality tests, social surveys, and logic games. The result? In 85% of cases, the AI made the same choices as the real person.
This level of accuracy could revolutionise multiple fields—from psychological research to market analysis, political forecasting, and even simulations of societal crises. Researchers could use AI-generated replicas to test public reactions to new policies, predict emerging trends, or determine which products are most likely to succeed—without ever having to survey real people.
When Machines Become More Human
According to the researchers, this AI technology goes far beyond traditional survey methods. While conventional questionnaires often provide only surface-level insights, these digital twins capture complex behavioural patterns. They consider not just facts but also emotions, beliefs, and social dynamics—something previously thought to be an exclusive domain of human consciousness.
However, the study also highlights the technology’s limitations. While the AI was highly accurate in replicating personality traits and social attitudes, it struggled with interactive decision-making scenarios. In economic simulations like the “Dictator Game,” which assesses fairness and social choices, the AI deviated more frequently from expected human responses.
Between Progress and the Risk of Manipulation
As fascinating as this technology is, it also comes with significant risks. Deepfake algorithms are already being used for fraud, identity theft, and disinformation. If AI models can convincingly simulate individuals, the potential for misuse grows exponentially. Fake interviews, politically charged statements, or even highly personalised misinformation campaigns could become frighteningly effective.
The researchers acknowledge these dangers but argue that AI simulations could serve as controlled ethical test environments. In theory, they could help us better understand human behaviour, anticipate social changes, and make more informed decisions.
The Future: A World of Digital Clones?
What if we all had our own AI replicas? A miniature version of ourselves, answering emails, making decisions, and even interacting with other AI clones? According to Joon Sung Park, the study’s lead author and a doctoral student at Stanford University, this might not be so far-fetched: “If you had multiple small versions of yourself running around, making decisions just as you would—that, I believe, is ultimately the future.”
Whether this vision becomes reality remains to be seen. But one thing is certain: with every AI advancement, the boundary between humans and machines becomes thinner. The question is no longer if artificial intelligence can imitate us—but what we choose to do with it.
Post picture: DALL-E3
