This piece is marked as syndicated from Business Insider, itself repeating Kaku himself on Big Think, so the very first step is to go there - and that quickly shows that they've omitted key qualifiers in Kaku's text:
But actually this experiment (the EPR experiment) has been done many times, and each time Einstein was wrong. Information does go faster than light, but Einstein has the last laugh. This is because the information that breaks the light barrier is random, and hence useless. (For example, let's say a friend always wears one red sock and one green sock. You don't know which leg wears which sock. If you suddenly see that one foot has a red sock, then you know instantly, faster than the speed of light, that the other sock is green. But this information is useless. You cannot send Morse code or usable information via red and green socks.)
Whether quantum particles "communicate" with each other is, if you really push it against the ropes, ultimately up to which intepretation of quantum mechanics you choose to believe. However, what's really clear is that even if they do communicate, they cannot communicate any causal information, so they cannot be used to communicate faster than light and cannot give rise to any time-travel paradoxes (which is the real worry with FTL communications). The details of how this comes about are subtle and there are plenty of questions on the subject on this site, but Kaku does a good job of explaining the limitations given the length of the piece.
However, given the inappropriate cutting they've done with the piece, I'll add this as well: as a rule of thumb, if Science Alert has content that is in direct contradiction with other media, it pays to slow down, read around, and think carefully about just how well they are representing the science. They have plenty of reasonable coverage, but they've graced these pages before in ways that didn't leave them looking very good. In this specific case, though, it just takes a bit of pulling on the thread to find the original source along with its more careful hedging of the conclusions and, moreover, pulling on that thread until you find that original source is a key bit of internet information hygiene that needs to be practiced pretty much always.