What follows is a very loose collection of thoughts in response to some of the points raised here.
1. Malik says: "I believe that no speech should be banned solely because of its content; I would distinguish ‘content-based’ regulation from ‘effects-based’ regulation and permit the prohibition only of speech that creates imminent danger."
Here is one worry with the above. Why should we want to prohibit only that speech which creates imminent danger? Strictly speaking, the only difference between imminent danger and other forms of danger is one of time. A pedestrian who is about to be hit by a car is in imminent danger. A person suffering from terminal cancer may not be. Presumably, neither harm is more significant than the other. If the "imminence" of danger matters, surely it is only because we have less time in which to avert the potential harm. So - it seems arbitrary to think that we may seek to prohibit speech that creates imminent danger but not speech that creates "non-imminent" danger.
Note also in the above quote that Malik explicitly concedes that speech can indeed create danger. It is important to bear this in mind, as Malik seems to contradict himself on this point later (see point 4, below).
2. Makik continues: "I oppose content-based bans both as a matter of principle and with a mind to the practical impact of such bans. Such laws are wrong in principle because free speech for everyone except bigots is not free speech at all. It is meaningless to defend the right of free expression for people with whose views we agree. The right to free speech only has political bite when we are forced to defend the rights of people with whose views we profoundly disagree."
Malik is in danger of creating a straw man here. Those who argue in favour of restrictions on hate speech are not arguing for "free speech for everyone except bigots". Rather, they are arguing for something like "free speech for everyone except people who express themselves in ways that are intended to incite hatred". It is possible to express views which are bigoted yet which are not intended to (or likely to) incite hatred. Most anti-hate speech laws include these kinds of requirements.
So, a defender of anti-hate speech legislation is certainly not committed to the claim that we should only protect freedom of expression for those with whom we agree. They are committed to the more restricted claim that freedom of expression should not extend to those whom intend to incite hatred. This represents only a small subset of disagreeable expression more generally (as does speech intended to incite violence). It is hardly meaningless to support freedom of expression except when it comes to inciting violence (as Malik himself does), so it is difficult to see why it would be meaningless to modestly expand that subset of prohibited speech to include incitement to hatred.
3. On the practical point, Malik claims: "And in practice, you cannot reduce or eliminate bigotry simply by banning it. You simply let the sentiments fester underground."
This may be true to an extent, but it misses the point - the aim of anti-hate speech legislation is not to reduce or eliminate bigotry, but rather to protect people from bigotry's harmful effects. Furthermore, it is implausible to think that we cannot reduce or eliminate bigotry if we prohibit hate speech. Malik offers the case of Britain in the 70's and 80's, but all that this shows is that anti-hate speech legislation is not sufficient to stop bigotry (but nobody claims otherwise).
4. Malik argues: "In blurring the distinction between speech and action, what is really being blurred is the idea of human agency and of moral responsibility. Because lurking underneath the argument is the idea that people respond like automata to words or images. But people are not like robots. They think and reason and act on their thoughts and reasoning. Words certainly have an impact on the real world, but that impact is mediated through human agency."
Malik has already accepted and goes on to reaffirm that people can be influenced by speech when that speech incites them to violence. To be consistent, then, Malik needs to reject prohibitions on incitement to violence (if he thinks this denies the agency of those who are incited) or to offer us a plausible reason for thinking that someone who we view as having been incited to violence does not have their agency undermined, but that someone who is incited to hatred does.
What we ought to do, I think, is to adopt the most sensible and straightforward position: People who incite (whether to violence or hatred) are performing acts which are intended to bring about certain kinds of harmful consequences. They are blameworthy to that extent. This does not require us to say that the person who is incited is blameless. If I supply you with a gun because I want you to use it to commit a crime and you do so, both of us are blameworthy.
5. Malik says: "Whether in London, New York, Berlin, or Kigali, speech should only be curtailed if such speech directly incites an act that causes or could cause physical harm to others and if individuals are in imminent danger of such harm because of those words."
It is worth highlighting this quote to show that Malik's view is extremely restrictive in terms of the kinds of harms that count, when it comes to legal prohibitions. For example:
(a) As mentioned previously, Malik's view limits us to protecting only against imminent danger (and this seems arbitrary).
(b) Malik's view is limited to physical harm. It takes no account of psychological harm.
(c) Malik's view is limited to discrete interactions between individual agents at particular (and limited) moments in time. It takes no account of long-term structural harms that can be created by a climate of bigotry, and which might manifest themselves in all sorts of complicated ways beyond the highly restricted cases Malik seems to have in mind.