Defy the data, assume the market is wrong. If people are unemployed, it's probably not because they can't do anything useful, but because the market hasn't found a way to use them yet. Their labor is undervalued, so buy their labor. At a severe slump, it will be economical for well off people to hire others to shine their shoes (or whatever the styles of 2025 dictate) just because they need the money. In your scenario, you know that R&D is underfunded, which likely means bright young people scavenging for less demanding jobs. Hire them to work on your AI or something with a faster/more likely payoff, or invest in an R&D company.