The AI model uses the data to create long patterns of recurring life events, categorizing and drawing connections between life events like income, education, or health factors. These connections serve as a foundation for the predictions the model makes, including a person's probability of mortality. However, the researchers note that these predictions are based on correlations, cultural and societal contexts, and existing biases in the dataset. They view their predictive model as the beginning of a conversation about how these tools should and shouldn't be used.
Key takeaways:
- Researchers have developed a new AI tool, life2vec, that can predict human lifespan and other life events using sequences of life events such as health history, education, job, and income.
- The AI model was trained on a data set from the entire population of Denmark, provided by the Danish government, and its predictions exceed the accuracy of current state-of-the-art models.
- Despite its predictive power, the researchers caution that the tool should not be used for prediction on real people, but rather as a foundation for future work, due to its reliance on a specific data set from a specific population.
- The researchers hope that their work can contribute to a more open, public understanding of how these predictive tools work, what they are capable of, and how they should and shouldn't be used.