I often read of "potential risks". This moved further into the realm of uncertainty with "can be a potential risk" in a recent, scientific magazine. Given that measurements of risk incorporate the less than 100% likelihood (which could be arbitrarily low) of an outcome occurring, what differentiates a "potential risk" from an actual one?
I've wondered this before, as well. I wasn't able to find anything on the internet that directly addressed this problem so I'm approaching intuitively. In my mind 'potential risk' is potential because the action is potential. If you are thinking about blowing the whistle on something illegal going on in your company, you could potentially be risking your career. The minute you do blow the whistle, you are risking your career. Until you actually go to the press you haven't taken the risk, so it's potential.
what differentiates a "potential risk" from an actual one
Adding 'potential' could indicate Knightian uncertainty (ignorance, unknowability), in addition to or instead of quantifiable risk. In other words, a 'potential risk' is one that is still unknown (could be zero probability), whereas a 'normal' risk has a known probability strictly greater than zero.
In most cases it would be redundant, but in the example provided by Waywardeevee, and in other examples of that character, I agree that it is not.
Oy gott. "Potential risk" and "potential danger" are grossly redundant. "Danger" denotes the possibility of harm. To say something is "potentially dangerous" means there is the possibility of the possibility of harm. To say "can be potential risk" means there's the possibility of the possibility of the possibility of harm. If this isn't the matruschka reality you're describing, just say "It's risky" or "It's dangerous."