Stanford and Google DeepMind Researchers Create AI that Can Replicate Human Personalities with Uncanny Accuracy
By interviewing 1,052 people from diverse backgrounds, researchers at Stanford and Google DeepMind have built what they call “simulation agents” – digital copies that can predict their human counterparts’ beliefs, attitudes, and behaviors with remarkable consistency.
Creating the Digital Copies
To create the digital copies, the team uses data from an “AI interviewer” designed to engage participants in natural conversation. The AI interviewer asks questions and generates personalized follow-up questions – an average of 82 per session – exploring everything from childhood memories to political views. Through these two-hour discussions, each participant generates detailed transcripts averaging 6,500 words.
Testing the Digital Copies
The researchers put their AI replicas through a battery of tests to assess whether they accurately copied various aspects of their human counterparts’ personalities.
General Social Survey
First, they used the General Social Survey – a measure of social attitudes that asks questions about everything from political views to religious beliefs. Here, the AI copies matched their human counterparts’ responses 85% of the time.
Big Five Personality Test
On the Big Five personality test, which measures traits like openness and conscientiousness through 44 different questions, the AI predictions aligned with human responses about 80% of the time. The system was superb at capturing traits like extraversion and neuroticism.
Economic Game Testing
Economic game testing revealed fascinating limitations, however. In the “Dictator Game,” where participants decide how to split money with others, the AI struggled to perfectly predict human generosity. In the “Trust Game,” which tests willingness to cooperate with others for mutual benefit, the digital copies only matched human choices about two-thirds of the time. This suggests that while AI can grasp our stated values, it still can’t fully capture the nuances of human social decision-making.
Real-World Experiments
The researchers also ran five classic social psychology experiments using their AI copies.
Experiment One: Blame and Intent
In one experiment testing how perceived intent affects blame, both humans and their AI copies showed similar patterns of assigning more blame when harmful actions seemed intentional.
Experiment Two: Fairness and Emotional Responses
Another experiment examined how fairness influences emotional responses, with AI copies accurately predicting human reactions to fair versus unfair treatment.
Easy AI Clones: What Are the Implications?
AI clones are big business, with Meta recently announcing plans to fill Facebook and Instagram with AI profiles that can create content and engage with users. TikTok has also jumped into the fray with its new “Symphony” suite of AI-powered creative tools, which includes digital avatars that can be used by brands and creators to produce localized content at scale.
The Future of AI Clones
Stanford and DeepMind’s research suggests that such digital replicas will become far more sophisticated – and easier to build and deploy at scale. “If you can have a bunch of small ‘yous’ running around and actually making the decisions that you would have made — that, I think, is ultimately the future,” lead researcher Joon Sung Park describes.
Conclusion
The research team acknowledges the potential risks of creating digital copies that are so convincing they can be used for malicious purposes. However, they believe that with clear consent from participants and strict data protection measures, AI clones can be used for good, such as supporting scientific research and improving public health messaging. As we push deeper into the uncharted territories of human-machine interaction, the long-term implications remain largely unknown.
FAQs
Q: What is the purpose of the research?
A: The purpose of the research is to create AI clones that can replicate human personalities with uncanny accuracy.
Q: How did the researchers create the digital copies?
A: The researchers used data from an “AI interviewer” designed to engage participants in natural conversation.
Q: How accurate are the digital copies?
A: The AI copies matched their human counterparts’ responses 85% of the time in the General Social Survey and 80% of the time in the Big Five personality test.
Q: What are the potential implications of creating digital copies?
A: The potential implications include supporting scientific research, improving public health messaging, and enabling more effective marketing and sales. However, there are also risks of malicious use, such as creating fake profiles or spreading misinformation.
