As AI is becoming more and more accurate and advanced in predictions, it is now on the cusp of predict things related to human behaviour as well.
For example, a Xeon-based supercomputer called Nautilus had predicted the location of the militant organisation al-Qaeda’s chief Osama Bin Laden within 125 miles, by collecting data dated back to 1945 from news clippings all over the world. The same supercomputer had also predicted the revolutionary uprising of Egypt that had occurred in 2011.
With this forecasting capability becoming more accurate, will it be possible to predict exact events in the world, for say, 10 years from now?
The Modelling Religion Project (MRP) has an international team of computer scientists, philosophers, religion scholars, and others collaborating regarding the same. They are building computer models that have several virtual people and their interactions. The “agents”, as they like to call these virtual people, interact with each other and with the environment around them. When they do this, their beliefs in regard to social aspects change.
Here’s How It Worked
The researchers use a detailed survey data of these attributes of a country and the agents are programmed to mimic the beliefs with respect to that country. The model was also trained to react to a set of empirically validated set of rules of social science about how humans tend to interact under various pressures. The model can predict these results after adding, say fifty thousand people and investing in one of the social policies, it tells how this artificial society will change.
The data used was between 1991 to 1998 and they tested with a list of different societal policies, taken from the International Social Survey Programme. The model was initialised from 1998 and was made to run till the year 2008.
Norway posed to be a good candidate to test this model because has gathered massive datasets of its population so far. It is also Syrian refugees and so it makes a good ground for this research. Predictions regarding what will happen to a particular city in Norway or even a particular neighbourhood, the model can be used.
Models Used To “Predict” The Future Of Religion
Using a model called Future of Religion and Secular Transitions (FOREST), it was theorised that there are four factors in which people tend to secularise:
- Existential security: It is a state of being secure enough to be taken as granted and you have enough money and food to survive. It is based on the view that religion arises with a lack of economic security.
- Personal freedom: It is the freedom to choose whether to believe in God or not.
- Pluralism: It is a system or a society with several different groups or states and principles and beliefs. It basically means a receptivity to diversity.
- Education: It is based on having some training in science and/or humanities.
These four factors boost secularism and an absence of either of these slows down the secularism rate. They believe that this is the reason why the US is secularizing at a pace slower than that of the Western and Northern Europe.
Another model called the Mutually Escalating Religious Violence (MERV). It aims to analyse what situations make xenophobic anxiety among religious groups. A very surprising thing is that the model could perfectly predict is the 2002 Gujarat riots.
The main aim of the MRP project was to provide politicians with a tool that will help them select the most effective policy option for a society to progress and cater. Which is indeed a good action, because if the leaders can know beforehand what policy is going to give a positive outcome, we might have a healthier, happier and more secure society to populate. By predicting such events, whether good or bad, would also make us have a chance to be able to take measures to avoid it, or overcome it.
How Reliable Are These Models?
With AI or statistical models, it often happens that you end up predicting something that you earlier did not intend or expect. Because these models are predicting something so crucial involving the society, whether they really are infallible or not is a very vital decision to inspect, for them to come into literal practical use in the society. Their prediction must be accurate enough for us to take actions on them.
Humans tend to seek the challenge of invading our own immortal creations; first religion and now perhaps AI sometime in the future. For centuries, theorists and scientists have been trying to find out religious beliefs and the compelling behaviour of religions. The MRP project greatly encapsulates the power of AI in social sciences and might soon be able to predict societal events of the future. Although the dependency of these models is currently still under investigation, as the world progresses, we might see AI taking over religion.
Try deep learning using MATLAB