The FCC and Congress (and for that matter, the ITU) persist in maintaining an artificial divide between the voice and data regulatory regimes that has not been logically sustainable since digitization of voice became widespread in the late 1990s. Perhaps the best expression of this inanity is Rep. Ed Markey's proposed amendment to the Freedom and Broadband Deployment Act of 2001 (the Tauzin-Dingell Bill):
The term 'high speed data service' includes any telecommunications service delivering data, represented by a string of binary bits transmitted digitally as a series of zeroes and ones, but such data services do not include, under any circumstances, bits representing voice communications, even though such voice services are also expressed digitally as a series of zeros and ones, data services being digital bits of special importance. Someone, preferably a senior corporate employee with years of experience in analyzing zeroes and ones, shall posses authority to determine which bits are which as they zip through a network literally at light speed. All decisions of such employee may be challenged and brought to the attention of government officials, but only after a wee bit, and only bit by bit.Rep. Markey's proposal was, sadly, ruled out of order by the committee chair. This amendment (which was clearly offered for comic value) underscores the fact that, at least in 2001, while telecommunication providers could not meaningfully distinguish between digitized voice data and other forms of data passing through their network, Congress (and therefore also the FCC) continued to operate under the assumption that they could. (For more on the Tauzin-Dingell bill, see this interesting article from llrx.com.)
Sadly, neither Congress nor the FCC appears to have learned anything in this regard since 2001; even as the FCC (and also Homeland Security) pushes commercial and public safety licensees into digital voice, the FCC continues to be burdened by an obsolete voice/data regulatory dichotomy. To be fair, much of this is forced on them by Congress, and even to some degree by the ITU (emission designators, which drive the regulatory picture at a very deep level, are based only on the superficial RF modulation technique and the ultimate payload, without regard to any intermediate encoding).
As to D*STAR itself, and its reliance on AMBE, I'll simply repeat what I said back in October: "We need to encourage the development of open digital voice standards, not make excuses for perpetuating existing closed ones just because they already exist."