Look at the OP news article, they took a picture - so they have already a transparent material (type of glass or plastic). There you go. My comment was about a low tech bigger scale endoscope instead of error prone high-tech robotics.
And glass and probably types of plastics can be quite hardened against negative effects of radiation. Think of transparent windows where you can look at the reactor core of research reactors, NASA space missions with high precision optics that last for decades, etc.
The issue in Japan is their culture, their problem with "loss of face" which inevitably caused problems in the hours when the worst case scenario unleashed - and the whole world watched 24h live while they did little and USA proposed (and had them ready on the nearby ship) to send in helicopters with backup generators but Japan made no official request for help - lesson learned, in the end US should have crossed their border and fixed the generators because Japan hasn't acted rational at all. Even the years later, they could have learned from Russia how to fix it quickly by sending in human people. But again false bride, and "loss of face" gets in their way. They keep going on with business as usual, and the leaks aren't fixed or anything worth mentioning. Chinese and Japanese have a related culture, yet Chinese know what they are doing and are very successful.
"All optical fibers undergo some darkening depending on a number of factors that include: ionization type, optical fiber core glass composition, operating wavelength, dose rate, total accumulated dose, temperature and power propagating through the core. Since attenuation is composition dependent, it is observed that fibers having pure silica cores and fluorine down doped claddings are amongst the most radiation hard fibers. To minimize damage consequences, it is better to use a pure silica core fiber at higher operating wavelength, higher temperature (accelerated recovery) and higher signal power (photo-bleaching)." -- https://en.wikipedia.org/wiki/Radiation_effects_on_optical_f... -- (shortened snip)