Yep; and let's not forget that it isn't our direct-to-consumer, inexpensive, genotyping microarray tests that are used as evidence in courts of law anyway. SNPs and presumed contiguous segments do not make for very precise individual identification. That comes from the autosomal STRs tested and cataloged in CODIS/NDIS and other, international databases.
The U.S. "DNA Identification Act" was established in 1994 and formalized use of the FBI's NDIS database, and also CODIS, the combined database of national, state, and local DNA references (yep; many large cities have their own databases that tie into CODIS, referred to collectively as LDIS). And it isn't just convicted felons whose DNA results are entered. Practices vary by state, but people not convicted--even exonerated--may have been swabbed and the results uploaded to CODIS; any number of "persons of interest" in a violent felony may have been swabbed and uploaded; DNA samples taken from a crime scene and sequenced may be entered whether the samples proved to be material to the case or not; and the most recent database in CODIS is one explicitly for missing persons: a push has been underway for a few years to make this as inclusive as possible, so the swabs done of immediate family members in a missing persons case may go into CODIS.
At the end of the day, the way law enforcement--or more accurately, folks like CeCe Moore--use our genealogy testing databases is no different than the way our Adoption Angels use them. They look for hints about who might be related to whom, and use that as a way to try to locate the "person of interest." Nobody has ever been, or will ever be, convicted of a crime based on our $59 SNP tests.