On 14 Feb 2018, at 20:49, John Horton <john.horton@legitscript.com> wrote:
Hmm, well, perhaps it's because I work for a company that processes quite a bit of data with a combination of algorithms and some human review, but I feel pretty confident that there are ways to simplify that with magic algorithms and forms.
Magic algorithms are fine in pattern detection because there is always a human review at some point or the cost of error is low, like in raising an abuse case that contains wording like supposedly", "allegedly" etc. In this case, every false negative comes with a tremendous liability. Also, if machine-learning technology and deep pockets for lawsuits become a requirement for being a registrar, you can count on the number of registrars dropping to single digits. Rubens