What will it cost to take existing boosters and man-rate them?

Man-rating is a red-herring. The Shuttle was never man-rated. The biggest thing that we can hope to achieve with man-rating is to develop malfunction detection systems that can give sufficient warning so the launch escape system can save the crew. We've done this before. The Redstone and Atlas boosters used for the Mercury program as well as the Titan II used for Gemini launches were missiles adapted for the purpose. By any realistic measure, the cost of developing and adding those detection systems to existing EELVs will be only a tiny fraction of the cost of developing the Ares I (which will require those systems as well).

Delta IVs and Atlas Vs are already built and flying. To date, they have a perfect track record. Each was designed to only require a small ground crew to prepare the boosters for launch and then to launch them. NASA is still talking about maintaining a "cast of thousands" for the Ares vehicles. That's one of the biggest differences between a government program and one designed to make a profit. In addition, using an EELV for the manned flights would increase their flight rates, lowering costs for all launches (manned or unmanned). I doubt the Ares I will ever fly more than 4 missions a year and the Ares V is unlikely to fly more than twice a year. That makes for very expensive launches.