Artificial intelligence (AI) has evolved rapidly over the past decade, touching nearly every aspect of our lives, from automating tasks to driving cars. Now, in a striking new development, AI technology has advanced to the point where it can predict something as profound as when a person might die—with an accuracy rate of 78%.
Introducing Life2vec: The AI That Predicts Lifespan
The innovative software, known as Life2vec, was developed by a team of experts from Denmark and the United States. The system uses a large language model (LLM), similar to popular AI tools like ChatGPT and Google's Bard, to analyse vast datasets. The researchers fed Life2vec data from Danish health and demographic records of six million individuals, encompassing a wide range of factors such as income, profession, and medical history.
This AI system operates by converting complex life data into narrative-style summaries. For instance, the system can analyse data points like, "In September 2012, Francisco received twenty thousand Danish kroner as a guard at a castle in Elsinore," to construct a comprehensive picture of an individual's life. This data is then used to calculate life expectancy with an impressive 78% accuracy.
Factors That Influence Life Expectancy
Life2vec's predictions are grounded in well-established research on factors that affect longevity. Previous studies have shown that certain characteristics—such as being male, smoking, or having a history of poor mental health—are associated with shorter lifespans. Conversely, higher income levels and holding leadership positions are often linked to longer lives.
The Danish dataset, used to train Life2vec, includes coded information for various life events and conditions, such as "S52" for a broken forearm or "IND4726" for working in a tobacco shop. These codes were converted into descriptive words, enabling the AI to build a detailed narrative of a person's life, including their profession, income, injuries, and even pregnancy history.
Testing Life2vec: How Accurate Is It?
The team tested Life2vec on a sample group of people aged between 35 and 65, half of whom had died between 2016 and 2020. The AI predicted with startling accuracy—78%—which individuals would die within that timeframe. Moreover, Life2vec proved to be more accurate than models specifically trained to predict the results of personality tests.
While this technology is groundbreaking, it is not without its limitations. The Life2vec model was trained exclusively on Danish data, which means its accuracy may not translate well to individuals from other countries. As Professor Sune Lehmann Jørgensen from the Technical University of Denmark, who led the research, pointed out, the model's predictions might not be as precise for non-Danish populations.
Ethical Concerns and Potential Misuse
One of the most pressing concerns surrounding Life2vec is the potential for misuse, particularly by big businesses such as insurance companies. Professor Jørgensen highlighted the ethical dilemma, stating, "Clearly, our model should not be used by an insurance company because the whole idea of insurance is that, by sharing the lack of knowledge of who is going to be the unlucky person struck by some incident, or death, or losing your backpack, we can kind of share this burden."
Despite these concerns, similar technologies are already emerging in the market. The European Union's AI Act, recently published in the Official Journal of the European Union, aims to regulate the development and deployment of AI technologies, ensuring they align with European values and fundamental rights. The Act is a comprehensive framework designed to promote human-centric AI, protect basic rights, and foster innovation. It also introduces stringent requirements for high-risk AI systems, mandates transparency, and establishes a robust governance structure to oversee AI activities within the EU.
With a two-year transition period set to begin shortly, companies, particularly those in the healthcare sector, must prepare to comply with this groundbreaking legislation. This is especially urgent for those who must address prohibitions on unacceptable risk AI systems within six months from August 2, 2024.
A Glimpse into the Future
Life2vec represents a significant leap forward in AI's capabilities, offering a glimpse into a future where AI could play a critical role in personal healthcare and life planning. However, as with any powerful technology, it brings with it a host of ethical and regulatory challenges that society must carefully navigate.
