Resist means without catastrophic corrosion here. We assume that there is no coating or anodic protection. The steel can be mirror polished and passivated. To make it more difficult, the shape is not a perfect sphere but a mechanism with some very bad water stagnation. Since all seas are different let's say that the part is a few miles away from San Francisco seashore. We are 50 meters underwater. There is no source of heat in the mechanism.
For this type of use and to make it simple we will calculate the Pitting resistance equivalent number. This is based upon the material composition.
PREN = 1 x %Cr + 3.3 ( %Mo + 0.5 x %W ) + 16 x %N
First of all, I would not test any steel that have a PREN inferior to 40.
Here we are left with Super Duplex, Hyper Duplex and Super Austenitic.
Then I would only choose steels made with very high end manufacturing processes in order to avoid any pollution. The heat treatment is also very important. Electropolishing can help to clean the surface.
Classical Super Austenitic include 904L, 254 SMO, AL-6XN, F44...
I prefer Duplex over Austenitic because it has better mechanical properties as long as you don't encounter very high temperatures.
Classical Super Duplex include F53, F55... These could be very interesting to test.
Super Duplex does not corrode easily. It usually happens at high temperatures or pressures. But let's say that there are very very bad conditions in this mechanism... We would be left with Hyper Duplex like 2707 or 3207. The good point of these steels is that these have the best mechanical characteristics available for this type of use.
I would say Thibd's answer excellent. However, one should look at the cost (including fabrication) on Monel, Incoloy 825, any Inconel or Hastelloy or Titanium . Although not "steel" they are metals that would answer the intent of the question. Also in "deep" water, corrosion of even carbon steel is not a problem because of low oxygen; but I think testing would be necessary to determine if the the subject location falls into that situation. And, although not exotic, attaching an aluminum , magnesium or zinc anode to a carbon steel object would certainly work and be the cheapest option. One would just need to check literature for the size and type of anode.