#digitaltwins — Public Fediverse posts

If it sounds crazy, maybe it really is. Here the craziness is expecting LLMs to behave like survey respondents.

"Scientists and practitioners are aggressively moving to deploy digital twins—large language model (LLM)-based models of individuals—across social science and policy research. We conducted 19 preregistered studies with 164 diverse outcomes (e.g., attitudes toward hiring algorithms and intention to share misinformation) and compared human responses with those of their digital twins (trained on each person’s previous answers to more than 500 questions). We establish an empirical benchmark for digital twin performance: Digital twins’ answers are only modestly more accurate than those from the (homogeneous) base LLM and correlate weakly with human responses (average correlation coefficient of 0.20). To guide future development, we document five ways in which digital twins distort human behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological biases, and (v) hyper-rationality. We make our full dataset and code public as a standardized testbed. Our results caution against premature deployment while laying the groundwork for the transparent, replicable, and iterative science necessary for responsible deployment of digital twins."

#SocialScience
#PolicyResearch
#DigitalTwins

https://www.science.org/doi/full/10.1126/sciadv.aeh8260?af=R

If it sounds crazy, maybe it really is. Here the craziness is expecting LLMs to behave like survey respondents.

"Scientists and practitioners are aggressively moving to deploy digital twins—large language model (LLM)-based models of individuals—across social science and policy research. We conducted 19 preregistered studies with 164 diverse outcomes (e.g., attitudes toward hiring algorithms and intention to share misinformation) and compared human responses with those of their digital twins (trained on each person’s previous answers to more than 500 questions). We establish an empirical benchmark for digital twin performance: Digital twins’ answers are only modestly more accurate than those from the (homogeneous) base LLM and correlate weakly with human responses (average correlation coefficient of 0.20). To guide future development, we document five ways in which digital twins distort human behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological biases, and (v) hyper-rationality. We make our full dataset and code public as a standardized testbed. Our results caution against premature deployment while laying the groundwork for the transparent, replicable, and iterative science necessary for responsible deployment of digital twins."

#SocialScience
#PolicyResearch
#DigitalTwins

https://www.science.org/doi/full/10.1126/sciadv.aeh8260?af=R

If it sounds crazy, maybe it really is. Here the craziness is expecting LLMs to behave like survey respondents.

"Scientists and practitioners are aggressively moving to deploy digital twins—large language model (LLM)-based models of individuals—across social science and policy research. We conducted 19 preregistered studies with 164 diverse outcomes (e.g., attitudes toward hiring algorithms and intention to share misinformation) and compared human responses with those of their digital twins (trained on each person’s previous answers to more than 500 questions). We establish an empirical benchmark for digital twin performance: Digital twins’ answers are only modestly more accurate than those from the (homogeneous) base LLM and correlate weakly with human responses (average correlation coefficient of 0.20). To guide future development, we document five ways in which digital twins distort human behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological biases, and (v) hyper-rationality. We make our full dataset and code public as a standardized testbed. Our results caution against premature deployment while laying the groundwork for the transparent, replicable, and iterative science necessary for responsible deployment of digital twins."

#SocialScience
#PolicyResearch
#DigitalTwins

https://www.science.org/doi/full/10.1126/sciadv.aeh8260?af=R

If it sounds crazy, maybe it really is. Here the craziness is expecting LLMs to behave like survey respondents.

"Scientists and practitioners are aggressively moving to deploy digital twins—large language model (LLM)-based models of individuals—across social science and policy research. We conducted 19 preregistered studies with 164 diverse outcomes (e.g., attitudes toward hiring algorithms and intention to share misinformation) and compared human responses with those of their digital twins (trained on each person’s previous answers to more than 500 questions). We establish an empirical benchmark for digital twin performance: Digital twins’ answers are only modestly more accurate than those from the (homogeneous) base LLM and correlate weakly with human responses (average correlation coefficient of 0.20). To guide future development, we document five ways in which digital twins distort human behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological biases, and (v) hyper-rationality. We make our full dataset and code public as a standardized testbed. Our results caution against premature deployment while laying the groundwork for the transparent, replicable, and iterative science necessary for responsible deployment of digital twins."

#SocialScience
#PolicyResearch
#DigitalTwins

https://www.science.org/doi/full/10.1126/sciadv.aeh8260?af=R

If it sounds crazy, maybe it really is. Here the craziness is expecting LLMs to behave like survey respondents.

"Scientists and practitioners are aggressively moving to deploy digital twins—large language model (LLM)-based models of individuals—across social science and policy research. We conducted 19 preregistered studies with 164 diverse outcomes (e.g., attitudes toward hiring algorithms and intention to share misinformation) and compared human responses with those of their digital twins (trained on each person’s previous answers to more than 500 questions). We establish an empirical benchmark for digital twin performance: Digital twins’ answers are only modestly more accurate than those from the (homogeneous) base LLM and correlate weakly with human responses (average correlation coefficient of 0.20). To guide future development, we document five ways in which digital twins distort human behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological biases, and (v) hyper-rationality. We make our full dataset and code public as a standardized testbed. Our results caution against premature deployment while laying the groundwork for the transparent, replicable, and iterative science necessary for responsible deployment of digital twins."

#SocialScience
#PolicyResearch
#DigitalTwins

https://www.science.org/doi/full/10.1126/sciadv.aeh8260?af=R

Michigan Daily: UMich study creates a digital AI copy of patient brain tumors to evaluate metabolic therapy potential. “A University of Michigan study published Jan. 6 used machine learning to create a digital twin replica of gliomas to shed light on the metabolic pathways activated by brain tumors. Gliomas — tumors composed of the nervous system’s glial tissue — are aggressive and complex.”

Forget waiting for replies! Eightfold co-founders just raised $35M for Viven, an AI digital twin startup that lets you 'query' your coworkers even when they're unavailable.

Is this a practical shortcut to productivity or just a new level of digital uncanny valley?

What's your take on AI digital twins for work?

#AI #FutureOfWork #TechNews #DigitalTwins #WorkTech

https://techcrunch.com/2025/10/15/eightfold-co-founders-raise-35m-for-viven-an-ai-digital-twin-startup-for-querying-unavailable-coworkers/

CS TECH Ai Reimagines India’s Tech Footprint with a Global Vision

#TycoonWorld #CSTechnology #CSTechnAi #DigitalTransformation #GeospatialIntelligence #MobilityEngineering #AIDrivenSolutions #SmartInfrastructure #DigitalTwins #SustainableDevelopment #SmartCitiesMission #JalJeevanMission #UrbanInnovation #EngineeringExcellence #AIforGood #InfrastructureInnovation #TechForImpact #GlobalTechLeadership #IntelligentInfrastructure

https://tycoonworld.in/cs-tech-ai-reimagines-indias-tech-footprint-with-a-global-vision/

Paul Cureton and I have just been informed that our book #Geodesign, #Urban #DigitalTwins and #Futures has gone to print today! This is very exciting, and for me at least a first! I am leading the tutorial content and managing the companion website content here (launch of site coincides with book release) : http://geodesigndigitaltwins.com/ .

In Nottingham (UK) today for 3 days teaching #cityengine to a lovely client. Invariably we will discuss #gameengines and #digitaltwins #geodesign and #3D #GIS 😀