Well, it depends on what you think the average human is capable of.
The way our society works is that you exchange your work for money in order to survive. Your whole survival depends on how much your work is valued on the labor market. If the only thing you are able to exchange is physical labor, congratulations, it's already worth almost nothing and you might starve in a near future! If the only thing you can exchange is prone to automation (hint: at the moment everything that is on the labor market is prone to automation), you will not make a dime.
So actually, what you are proposing is that people should evolve skills which value is not going to sink due to automation. Which begs two fundamental questions:
1. What are those skills? You speak of information, but automated information generation is already successful. Automated creation is on the rise, and partly already successful. Or you mean we will discover that soon enough, and the people in college right now are certainly not being taught those next things.
2. Do you really think everybody can catch up? There's the trend: the jobs that remain are the highest skilled one, which by definition are reserved to the lucky few that are able to do them. I don't think 90% of the population can defend a PhD, whatever the domain they prefer. Maybe you think that 90% of the population can improve their skills so as to beat the machine, forever.
So it all boils down to this: your faith in what the average human is capable of that is not automatable and has value on the market. Long term trend, I think is nil. Short term trend, I think it's below 30% of the population.