In a groundbreaking revelation, artificial intelligence (AI) has demonstrated the ability to predict future events in people’s lives, including estimating the time of death. A collaborative research project from DTU, University of Copenhagen, ITU, and Northeastern University in the US has harnessed the power of transformer models, similar to ChatGPT, to analyze vast amounts of data about individuals. The resulting model, named Life2vec, has proven to outperform other advanced neural networks, raising both excitement and ethical concerns.
Also Read: Get Tomorrow’s News, Today: NewsGPT Introduces New AI for ‘News Forecast’
The Power of Language-Driven AI
Artificial intelligence, specifically designed to model written language, has shown promise in predicting life events. The research, outlined in the article “Using Sequences of Life-events to Predict Human Lives,” delves into the capabilities of transformer models. These models, like ChatGPT, process language to systematically organize data and forecast various aspects of a person’s life.
Life2vec Model Outperforms
After an initial training phase where the model learns patterns from extensive data, Life2vec has demonstrated remarkable accuracy in predicting outcomes. These predictions include personality traits and the dates and times of certain life events. The researchers, led by Sune Lehmann, a professor at DTU, express their scientific fascination with the precision of the model’s answers and the underlying data aspects that enable such accuracy.
The AI Death Calculator
Of all the predictions made by Life2vec, the scariest one has been of a person’s death. The team tested Life2vec in a group of people aged 35 – 65, half of whom died between 2016 and 2020. The model, now known as “the AI death calculator” managed to predict their death with an accuracy of 78%. The newly developed AI is still being trained and tested to improve its accuracy in the prediction of life expectancy, amongst other major events.
Decoding Life Sequences
Life2vec encodes data in a comprehensive system of vectors. More like a mathematical structure organizing information about an individual’s time of birth, schooling, education, salary, housing, and health. The researchers link human life to a sequence of events, akin to how a language sentence consists of a series of words. This innovative application of transformer models provides a new perspective on analyzing what the researchers term “life sequences.”
Ethical Dilemmas Surrounding Life2vec
While the potential of Life2vec is groundbreaking, ethical questions loom large. Concerns include safeguarding sensitive data, ensuring privacy, and addressing biases in the data. The researchers emphasize the need for a deeper understanding of these challenges before deploying the model for assessing an individual’s risk of contracting diseases or other preventable life events.
Also Read: OpenAI Prepares for Ethical and Responsible AI
A Democratic Conversation on Technology
The article’s authors assert that the discussion around Life2vec’s implications should become part of a democratic conversation. They highlight the parallel technologies already used by tech companies, which track behavior on social networks, create highly accurate profiles, and predict human behavior. They argue it is crucial to collectively consider the trajectory of technology & evaluate whether this development aligns with societal values.
Our Say
The integration of AI into predicting future life events opens up a realm of possibilities and challenges. As Life2vec evolves, addressing ethical concerns becomes paramount. The delicate balance between technological advancement and protecting individual rights must be maintained. The researchers advocate for a democratic discourse to shape the future of such transformative technologies. This will hopefully ensure AI development aligns with the values and aspirations of society.
