The authors conduct an ex ante policy evaluation of the Trump Administration’s public charge rule using artificial intelligence to recreate immigration official decision-making. The analysis concludes that the number of immigrants who should be evaluated as likely to become a public charge under the rule is zero.