I know people have been looking into MBL and quantum memories, the problem seems to be that symmetries you'd need for topological phases either wreck MBL (beyond like SU(2)) or are the result of some quantum integrability (which also is not good for MBL)

source: my advisor works on passive error correction