You give the control arm the current best standard of care until and unless it becomes statistically obvious that your test arm is better and then ethically you're supposed to switch everyone to the drug at that point. It's not like the control group is denied the best care we know how to provide, just to accommodate experimentation.

This is an objection that could be raised against literally any clinical trial.

If we knew that a treatment was effective and safe, it would indeed be immoral to withhold it from people. But we don't know that. Finding that out is the whole point of having trials!

What if the treatment causes adverse effects? That way the control population are better off.

The alternative is to treat everyone as the null hypothesis population, or administer the treatment to everyone and then be unable to determine if it's effective.

Both of these are strictly worse, since neither of them can result in the discovery of an effective treatment.

They might be more likely to die if you give them the wrong medicine.