Many observers immediately labeled Fox's block a violation of the principle of "network neutrality"—the idea that Internet service providers should allow subscribers to access all legal content online. Neutrality rules have been the subject of fierce debate in Washington, and activists are constantly on the lookout for perceived anti-neutrality maneuvering.
If Fox's move violated "neutrality," though, it wasn't in the way we've long defined that term. Advocates for net neutrality rules have mainly been concerned about the power that cable and phone companies can exert on the Internet. The theory is that in most local areas, broadband companies exist as monopolies or duopolies—you can get the Internet from your phone company or your cable company—and, therefore, are in a position to influence online content. What if, for instance, AT&T demanded that YouTube pay a surcharge every time a customer watches a video? To prevent such abuses, the Federal Communications Commission imposed Internet "openness" guidelines (PDF) in 2005, and since then regulators and lawmakers have been arguing about how to make those guidelines both permanent and enforceable.
But this Fox-Cablevision-Hulu scenario turns the neutrality debate on its head. Here, it wasn't the broadband company—Cablevision—that blocked customers' access to content. Instead, it was the content company, Fox, that imposed the ban. Why is that distinction important? Because while it's easy to think of justifications for imposing neutrality regulations on broadband companies, it's less clear how we should feel about imposing rules on content providers. Telecom companies are regulated by the FCC, and there's a long history of the government forcing "openness" rules on public communications infrastructure. If the government can prohibit phone companies from deciding whom you can and can't call, shouldn't we have a similar rule preventing ISPs from deciding what you can get on the Web?
"First, the artificial intelligence (AI) algorithms had to train themselves. They used about 78% of the data-some 295,267 records-to search for patterns and build their own internal "guidelines." They then tested themselves on the remaining records. Using record data available in 2005, they predicted which patients would have their first cardiovascular event over the next 10 years, and checked the guesses against the 2015 records. Unlike the ACC/AHA guidelines, the machine-learning methods were allowed to take into account 22 more data points, including ethnicity, arthritis, and kidney disease."
