The autosomal tests, as done by AncestryDNA, will in theory pick up even cases of chimera because the sources of the chimera were either the same or related parents, and the genotyping will create some strange results.

Chimeras become more confusing when dealing with heteroplasmy, which is a problem sometimes in mtDNA tests.

Often news stories do not adequately explain "DNA tests" and just throw out that phrase, not explaining what kind of DNA tests were undertaken. Again, mtDNA is a case that is known for heteroplasmy but AncestryDNA is not an mtDNA test.

Unusual cases are usual flagged by the labs when dealing with these autosomal tests (such as polysomy of the sex chromosomes.) If a case of chimerism throws off the genotyping then the lab will indicate either a failed test or (at least in the case of 23andMe, but I think not AncestryDNA) the company will ask the tester to contact them.

Regardless, these cases are so exceedingly rare it is not productive to pursue these as explanations for people who receive successful test results. If two people do these autosomal tests (like AncestryDNA) and do not share statistically significant half-identical regions then they do not share a recent (within 4 generations) common ancestor.