Those who regularly read Tyler Cowen’s and Alex Tabarrok’s Marginal Revolution -blog, like me, that Tyler is a big fan of artificial intelligence (AI). Partly because of his messages and partly because of praising reviews by friends on Facebook, I realize that I have to use it more.
That said, I want to comment on one Recent message from Tyler in which he was linked to one analysisDone by OpenAi’s Deep Research Tool, of the costs and benefits of USAID. This is what Tyler asked to do:
What are the best sources to read about American help and the costs and benefits? I want serious analyzes, based on evidence, data and possibly economic models. How are the program in cost-benefits for terms? Please try to look beyond the rhetoric on both sides of the debate, Pro and Con, and come to an actual assessment of the office in net cost-benefits. A paper of five to ten pages should be good in each style, with full quotes.
Note that Tyler asked to ‘look beyond the rhetoric’. Since rhetoric, as Deirdre McCloskey often reminds us, the art of effective or convincing Speaking or writing, that is, the art of arguing, “It is not logical to ask AI to prevent rhetoric. Preventing it is to prevent them from arguing. Assessing costs and benefits are arguing.
Perhaps, however, Tyler expects AI to have the same wrong idea about what rhetoric is that the majority of the audience has. So maybe it’s no problem. Although I think it is; See below.
This is what I noticed. The answer of deep research is An argument. And not only an argument, but also one that is somewhat one -sided. Here is, for example, how the idea can be made with some of the subsidies and interventions of USAID:
Democracy and stability: The absence of USAID’s democracy programs is more difficult to play because changes in Governance Pad are dependent. In some cases, local forces could have dominated for democracy, even without external help (eg Eastern European desire to become a member of the EU was a strong motivator). However, it is likely that the progress would have been slower. Without technical support for elections and civil society, emerging democracies can falter or have seen more disputed processes. In places like Kenya in 2013For example, American support for election committees and peace building helped prevent violence; Without that there could have been a repeat of the violence after the 2007 elections. On the other hand, it could be said that in certain countries the absence of political help of the US could have reduced the suspicion of foreign influence and led to more organic change (a point that increases critics, although evidence anyway is scarce). In general, the counterfactual suggests that The world would not be more democratic if USAID was never involved – In fact, some profit would probably be absent in freedom and rights.
Note that it does not discuss the idea that USAID may have been used to overthrow governments. I don’t know if it is true that USAID money was used to overthrow the government of Bangladesh. This part in The Times of India says it might be true. But note that deep research does not even raise the problem.
Dealing with other problems is similar. It takes a complaint against USAID, vaguely suggests how it could be true and then says that things improve.
It also literally does not mention part of the abuse of the money that the doge people have marked. Perhaps the direction to prevent rhetoric was taken as a direction to prevent criticism from being mentioned for which the critics declared their case passionately. So maybe Tyler should not have asked to prevent rhetoric.
I am not saying that the deep research approach is completely wrong. I just point to the limits and express my skepticism. To his honor, Tyler’s mention of other sources means that he does not do deep research as the last word on this subject.