Other than it being in violation of NEC (National Electrical Code), it will probably work if you have some spacing between the AC wires and the ethernet cables, and if the AC wires are parallel, as in Romex. Loose wires in conduit are going to be a problem.
The failure mode is 60 Hz pickup by the ethernet transceiver. Most of the 60Hz will be rejected by the intentionally lousy frequency response of the typical ethernet transformer. The CAT5 cable and ethernet input circuitry also has very good common mode rejection, which is mostly what's keeping stray signals out of your ethernet transceiver. Under ideal conditions, the amount of AC pickup by the CAT5 wiring should be identical in each wire and therefore cancel.
That's fine if there is some distance between the AC wiring and the CAT5. Past about 4 wire diameters, the pickup on each wire is essentially identical and your ethernet will work as advertised. However, when the AC wire are at different distances from the CAT5, you're going to get different signal levels on each CAT5 wire. Since they're different levels, the common mode rejection of the input xformer can't get rid of the crap.
The worst case situation happens when the wires are right next to each other. Although the wire spacing is small, the differences in wire spacing is large. Therefore the induced voltages are large.
Realistically, it takes a considerable length of parallel run AC and CAT5 before there will be any noticeable problems. My guess(tm) is that the problems start at about 25 meters with loose wires. Romex will probably work over a longer distance.
The real danger is safety from electrocution and induced power line glitches. I'm sure you put the convenience of sharing the conduit, over having your insurance company invalidate your fire insurance and the potential risk of electrocution. Therefore, I won't suggest that perhaps the NFPA has a clue and that your safety might be worth considering.