What approaches are there to apply AI to global economic processes?


Today, AI is mainly driven by own-profit-oriented companies (e.g. Facebook, Amazon, Google). Admittedly, there's a lot of AI in the health sector (even in the public health sector) and there's a lot of AI in the sustainability sector – but also mostly driven by obviously own-profit-oriented companies (e.g. Tesla, Uber, Google).

On the other side, one often hears from hard-core economists that centrally planned (= public-profit-oriented) economies (or economic principles) are "the work of the devil" - and that they failed all over history (sometimes for understandable reasons).

But intelligently planning global economic processes and applying these plans with the help of state-of-the-art AI - given the huge amounts of really big data available, and given the argument that globalization is finally for the benefit of all - would seem to be a rewarding endeavour, at least for parts of the AI community.

Why isn't this endeavour undertaken more decidedly? (Or is it?)

Where do I find approaches to apply AI to global economic processes? (Not only describing and understanding but mainly planning and executing?)

Hans-Peter Stricker

Posted 2019-05-02T18:44:59.150

Reputation: 391



It's currently just too complex

The different sources of information are too varied, in economics this is often referred to as a local knowledge problem, which hampers many large scale plans. Humans can react to slight differences like respecting local traditions, landscapes, history but an artificial intelligence would (currently at least) struggle not to generalise over such a large scale as a whole country's economy.

The real work in this case (and actually most 'AI' tasks) would be collecting all the necessary data. Here that part job is currently insurmountable.

Currently a human lead planned economy would do a better job.

Lio Elbammalf

Posted 2019-05-02T18:44:59.150

Reputation: 256

i feel the data exists to generate a decent simulator, and by extension model, the issue is as you said its complexity-- it would require heavy resources and man-hours for research which the government probably wouldnt want to spend on a non-gaurantee... but they do spend alot on smaller more compact research focuses, so who knows, one day the want, the resources, and the execution can all come together – mshlis – 2019-06-04T14:12:48.230

@mshlis Not saying you're wrong but I would be surprised if the data exists with the level of detail needed to make predictions to a better than human level. Do you know for sure this data exists? – Lio Elbammalf – 2019-06-04T14:31:25.833

nope not at all, but my speculation arises from the belief that the data exists but is not collected/organized. There are endless case studies, economic datasets, etc available, and if parsed and preprocessed could provide information that is similar to the concept at hand. My comment about generating a 'decent simulator' is probably nieve, especially because validation would be difficult, but i dont think it unfeasible – mshlis – 2019-06-04T15:13:58.400

@mshlis We do have a lot of data but I think the key point here is the level of work needed to make predictions at a level better than we could with a human planned economy. The amount of local knowledge needed for this level of economic planning requires a lots more ground work before we can start building usable models and making predictions. – Lio Elbammalf – 2019-06-05T14:02:44.477

Are you sure that it is just currently too complex and not in principle unachievable? That means: Is this your personal opinion? – Hans-Peter Stricker – 2019-06-12T14:01:28.517


I found this article from James Manyika (MacKinsey) helpful: Applying artificial intelligence for social good

Hans-Peter Stricker

Posted 2019-05-02T18:44:59.150

Reputation: 391