Sometimes there's no obvious solution to a problem and all you can do is call attention to the problem. This is definitely one of those cases: The issue isn't that there's some obvious source of oversight that was disregarded - the issue is that dangerous and potentially harmful experimentation is carried out without oversight. Maybe sufficient oversight is impossible in these scenarios, but if so it should be an informed decision made after long, thoughtful consideration, probably at something approaching an industry level.
First, the article mentions that Facebook does have multiple levels of internal oversight. I really don't think this kind of research can be externally regulated, and attempting to do so would just discourage publication of the findings. Businesses do controlled experiments on their customers all the time, with the most relevant example being A/B testing. The danger is extremely limited, and it seems like overreaching to have some sort of international IRB for all experiments that online businesses may need to conduct to optimize the content displayed to their visitors. Publishing research is a good thing.
Quotes in the article from research team employees suggested the opposite: that experiments were often run without anyone else on the team even knowing about it. That doesn't sound like oversight to me.
