Let's say that people think MD5 is a good cryptographic hash function. The worst single possible thing you can do is ban-hammer the promoters of this idea away. This is extremely dumb "Kill The Messenger" thinking, the promoters of those idea are exposing real communication problems (namely, the use of "hash function" to denote roughly similar concepts but are also very different in very important details). What you do, if you really think this Md5-is-cryptographically-secure problem is serious, is that you treat the success of the promoters of this false idea as a measure for how hard you failed to communicate basic cryptography to software devs. Do a modification to your education\media ecosystem, then measure how successful the false prophets are, if their success goes up or stays constant then you still haven't addressed the real problem, if it goes down then you're successfully fighting "misinformation", repeat till you drive their popularity to the ground.