We should be long past the point technologically where we could feed every available data point related to the global economy into a computer ... run it through an algorithm, and see what result.

I'm sure that something like this already exists at a very high level in relation to the global banking, finance and insurance market. Put in place after the 2008 financial crisis, the system is designed warn 'regulators' of the potential for another global financial crisis.

Is anybody working in this field?

How far are we along with this type of computer modeling?

Any predictions on what the future will hold for this type of thing when we make the jump to quantum computing (if we haven't already)?

My wife does data science for a financial company.

It's not as easy as it sounds of just inputting numbers and some algorithm outputs what you want.
Too many variables to account for.

Read up on Ray Dalio and Bridgewater.

Here are several interesting podcasts that talk directly about AI or indirectly. Even an ETF that uses AI to do its selections.

https://www.theinvestorspodcast.com/episodes/tip158-artificial-intelligence-rise-of-robots-martin-ford/

https://www.theinvestorspodcast.com/episodes/artificial-intelligence-finance-wes-gray/

https://www.theinvestorspodcast.com/episodes/tip190-jim-rickards-part-1-central-banking-taxes-and-crypto/

https://www.theinvestorspodcast.com/episodes/tip191-jim-rickards-part-2-ai-global-finance-and-crypto/

https://www.theinvestorspodcast.com/episodes/superforecasting-summary/

Stat Monitor Repairman said:

---

I'm sure that something like this already exists at a very high level in relation to the global banking, finance and insurance market. Put in place after the 2008 financial crisis, the system is designed warn 'regulators' of the potential for another global financial crisis.

---

Stress testing. Existed far before the financial crash and still does.

Stat Monitor Repairman said:

---

We should be long past the point technologically where we could feed every available data point related to the global economy into a computer ... run it through an algorithm, and see what result.

I'm sure that something like this already exists at a very high level in relation to the global banking, finance and insurance market. Put in place after the 2008 financial crisis, the system is designed warn 'regulators' of the potential for another global financial crisis.

Is anybody working in this field?

How far are we along with this type of computer modeling?

Any predictions on what the future will hold for this type of thing when we make the jump to quantum computing (if we haven't already)?

---

Your enlightenment hubris is charming... Or at least it would have been at Tomorrow Land in 1965.

What public companies are the most at the forefront of this?

GE said:

---

What public companies are the most at the forefront of this?

---

It would take world scale compute to answer the OP - we're not there yet.

AWS and other cloud compute are great enablers of ML for public use.

There's a pretty wide open set of algorithms available to the public on all the major platforms, depending on the industry these algos/methods will be adapted to the data product intended for use (e.g. Medical image analysis or oil reservoir analysis). If a truly unique method is created they'll patent it. So, there will be leaders in adapting the tech in each sector.

Most major companies have already launched and hold their own data science groups for this reason.

https://en.wikipedia.org/wiki/Black_swan_theory

https://strategiccfo.com/black-swan-events/

This is what makes modeling stuff like that hard

Ranger is correct.

Even if you did have good input data, we don't have enough output today (the economic ups and downs) to train a robust AI model.

What you are asking for is more regression analysis, which is widely used in economics and finance.

You know how most people ridicule social science for being not science? Well it just turns out that modeling human behavior is probably the most difficult problem in all of the sciences. Any of our models, theories, and data on it today are woefully incapable of even consistently explaining the behavior or motivations of large groups, let alone individuals. Further, machine learning wouldn't magically be able to conduct the experimental research necessary to begin filling the gaps anyways due to the ethical prohibitions on that.

There is too much data and not enough humans to input it or processor power to analyze it.

What data gets fed into this machine if it existed in 2005.... the bogus AAA bond ratings? Or the actual mortgage default rates? Does the important data that could be predictive even being collected and monitored?

You also over estimate machine learning predictive ability.

It cannot predict things when presented with novel data it hasn't been trained on. This is why Tesla's have driven into the side of semi trailers, because it hadn't been trained on that scenario, therefore it couldn't identify the danger.

The same applies to your hypothetical economy predictor. The events that cause booms and busts are often novel events, aka black swans. The side of a semi truck was a black swan for Tesla's algorithm.

I don't know about machine learning but the OP has me itching to read the Foundation Series by Isaac Asimov again.

Might be time to revisit this issue.

We are a breakthrough in quantum computing away from all hell breaking loose on this front.

I think it's about to hit us before we realize what's happening.