Predicting the future is a task of philosophy, religion and Fortunetelling magic for centuries. Today, with the development of artificial intelligence (AI), the question of the capabilities of the machine for the future events of Foresee, is acquired by scientific sensitivity. Who can really look at the future, or its ability is limited by past analysis? Rambler will answer this question.
What does it mean when predicting the future?
In strict sense, who does not predict the future in the metaphysical sense – it builds probability forecasts based on large data arrays, using machine models and statistics. This can be compared to the way meteorologists predict the weather: accuracy increases according to the usability and structure of the data, but never achieved complete certainty.
Predict in probability
AI's modern models have demonstrated the ability to predict behavior, economic and even biological phenomena. For example, in the field of medicine, neurological networks may predict the risk of stroke or cancer with superior accuracy compared to human experts, writing Natural Magazine. Such models are analyzed by thousands of medical images, medical history, genetic data, identifying models that cannot be accessed to the human eye.
Can create a person's digital copy
Forecast in social systems
Everyone finds applications in politics and economy. Learn Google DeepMind It shows that an agent with multitasking training can predict the motivation of the market and change in user behavior models with high accuracy.
However, such reflexive social systems – they change under the influence of the predictions themselves. As economist George Soros noted in the theory of reflexes, an effort to predict itself becomes a factor of changes. This effect limits the accuracy of whom in predicting social future.
Predict the future scientific discoveries
One of the most attractive examples is the GPT-4 model, capable of giving new scientific hypotheses, based on existing documents. Learn Cornell University In 2024, it shows that AI has successfully proposed a number of new relationships between biological marks in cancer, some of which have been confirmed by experimental.
Limitations: Entropy and Chaos
However, basic physical restrictions, such as Entropy and chaotic dynamics, creates barriers that cannot be overcome for absolute predictions. In systems with high sensitivity to the initial conditions (butterfly effects), even the smallest deviations cannot be forecast long term.
Moreover, the quantum is not guaranteed at the level of basic particles (Heisenberg principles) that limit the accuracy of the most ideal models. As noted in research Cornell University For quantum calculations, AI can only evaluate the probability, but do not make a determined forecast under the conditions that do not guarantee quantum.
In short, AI is capable of forecasting, but not predicted in a mysterious sense. Its power is in handling huge data arrays, identifying correlations and building probability models. However, he has no intuition, a warning or a knowledgeable in appearance as a person.
Earlier, we wrote how Robotov taught reading the emotions of everyone.