It's true that in principle if a falling observer was emitting light in a completely continuous way as in classical electromagnetism (rather than in discrete photons as in quantum electrodynamics), and outside observers could detect light waves that had been redshifted to arbitrarily large wavelengths (even larger than the diameter of the observable universe, say), then an outside observer could still see the image of a falling observer suspended above the horizon forever. In practice, because light is emitted in discrete photons, and because radio waves of very large wavelengths are too difficult to detect, an outside observer will cease to see any light from an infalling observer after they've gotten close enough to the horizon, as explained in the "Won't it take forever for you to fall in?" section in this answer from the Usenet Physics FAQ. Also, if you read the "Will you see the universe end?" section in that answer, you'll see that if Cooper were watching the outside universe as he approached the horizon, he would not see an infinite number of years pass on outside clocks as he approached the horizon, and he'd still see everyone outside aging at a finite rate after crossing the (outer) event horizon (a rotating black hole is predicted in relativity to have a second inner Cauchy horizon where an observer falling through would see the entire history of the universe before crossing it, but as described in The Science of Interstellar, Cooper was rescued by the Tesseract before reaching this horizon).
But even if they could still see the image of Cooper very close to the horizon, I don't think there would be anything paradoxical about them also seeing a Cooper who had escaped from the black hole. The reason is that the Tesseract seems to have scooped him out of one region of spacetime and deposited him in another region that would have been impossible for him to reach from the first region if not for the Tesseract (there's no normal way to escape from the interior of a black hole back to the outside). As an analogy, suppose there was a traversable wormhole with one mouth 1 light-year from Earth, and another mouth 5 light-years away. Suppose that in 2015 an alien who has been parked near the mouth 5 light years away decides to dive into it, emerges from the mouth 1 light year away on the same date in 2015, then travels at half the speed of light to Earth, arriving in 2017. Then we will meet the alien in 2017, but if we point sufficiently powerful telescopes at the mouth that's 5 light-years away, then since light signals from that distance take 5 years to reach us, we'll be seeing that mouth as it was in 2012, and so we should still be able to see the alien in his ship parked next to it. There's nothing really paradoxical about this, it's just a consequence of the fact that the light we're seeing from the alien 5 light-years away traveled to us the "long way", while the alien we see on Earth took a shortcut through spacetime.
Similarly, if someone dived into a black hole, and once inside found a mouth of wormhole whose other mouth was outside the black hole, so that the falling observer could escape and meet up with another observer orbiting outside the black hole, the outside observer might still be able to see a delayed image of the falling observer near the horizon* while also seeing a flesh-and-blood version of the same person next to them. The Tesseract probably provides a similar "shortcut" linking different regions of spacetime that wouldn't be reachable from one another otherwise.
*Technically the existence of the wormhole would change the location of the absolute horizon, which is defined as the region of spacetime from which it's impossible to escape and avoid the singularity, but there would still be an apparent horizon that would have the same measurable properties as the horizon of a similar black hole with no wormhole inside it.
