Brother
Professional
- Messages
- 2,590
- Reaction score
- 539
- Points
- 113
Is it wise to trust machines with the secrets of human destinies?
A joint study conducted by scientists from Denmark and the United States showed that the use of artificial intelligence based on transformer models, similar to ChatGPT, allows you to predict various events in people's lives, including even the date of their death.
In a new paper titled "Using Life Event Sequences to Predict Human Lives," published in the journal Nature, researchers analyzed data on the health and labor market relationships of 6 million Danes in an AI model dubbed Life2vec.
After initial training, the model demonstrated the ability to outperform any other advanced neural network, predicting outcomes such as people's personality and time of death with high accuracy.
Professor Sune Lehmann of the Danish University of Technology stressed that it is not the prediction itself that is exciting, but the data aspects that allow the model to give accurate answers to questions about future events based on past conditions and events.
The Life2vec model uses vectors to encode data about time of birth, education, salary, housing, and health. This data helps you effectively predict a wide variety of events with high accuracy.
Despite the huge potential of their creation, the researchers point out ethical issues related to the protection of sensitive data, privacy, and the role of possible model bias. These issues need to be worked through in depth before Life2vec can be used to assess risks, such as diseases or other preventable life events.
Professor Lehmann also stresses the importance of including this topic in the democratic dialogue, given that similar technologies are already being used by technology companies to predict human behavior.
The researchers next step is to integrate other types of information, such as text and images or information about social connections, which opens up new perspectives on the interaction of social and health sciences.
A joint study conducted by scientists from Denmark and the United States showed that the use of artificial intelligence based on transformer models, similar to ChatGPT, allows you to predict various events in people's lives, including even the date of their death.
In a new paper titled "Using Life Event Sequences to Predict Human Lives," published in the journal Nature, researchers analyzed data on the health and labor market relationships of 6 million Danes in an AI model dubbed Life2vec.
After initial training, the model demonstrated the ability to outperform any other advanced neural network, predicting outcomes such as people's personality and time of death with high accuracy.
Professor Sune Lehmann of the Danish University of Technology stressed that it is not the prediction itself that is exciting, but the data aspects that allow the model to give accurate answers to questions about future events based on past conditions and events.
The Life2vec model uses vectors to encode data about time of birth, education, salary, housing, and health. This data helps you effectively predict a wide variety of events with high accuracy.
Despite the huge potential of their creation, the researchers point out ethical issues related to the protection of sensitive data, privacy, and the role of possible model bias. These issues need to be worked through in depth before Life2vec can be used to assess risks, such as diseases or other preventable life events.
Professor Lehmann also stresses the importance of including this topic in the democratic dialogue, given that similar technologies are already being used by technology companies to predict human behavior.
The researchers next step is to integrate other types of information, such as text and images or information about social connections, which opens up new perspectives on the interaction of social and health sciences.
