We analyze how AI reshapes the investment, organization and outcome of scientific production. Using a corpus of both funded and unfunded proposals to a large international funding agency, as well as detailed budget information, we show that AI primarily has changed the inputs and organization (and not the output) of research, lowering the cost, enlarging the teams, and shifting resources from equipment to human capital. Link to pre-print: https://arxiv.org/abs/2603.27956
Read More