Crowdsourcing is a good option to consider, but only once professional people have gone through the sidescan sonar material to no avail. Keeping the data hidden after official investigators have given up wil only inspire conspiracy theorists. But lessons from the Fosset crowdsourced search must be taken in account:
1. Users need to trained before giving access to data - make them go through a online course of identifying various objects from sonar scan results.
2. Have *all* data available - debris findings and time, sound recordings from phase 1, etc.. Some people might find creative ways of combining data.
3. Make the reporting path non-intrusive. Instead of calling the investigators just click online of objects of possible interest, write a report of what you think you are seeing and let others comment on the credibility of finding.
Even if the crowd doesn't find AF447, one would presume something of scientific interest would be found. After all these seabeds are still not very well studied.