The author of this article is bringing up a valid point, but also focusing on personal characteristics that are likely to further incite the harassment they want to stop. Some people love to get under the skin of other people, and this article is telling them where to focus. The problem of gaming harassment is not focused on or unique to a specific gender or race -- we all deal with it to some degree. I'd argue that skill and age play just as big a role.
Twitch, XBox Live, Steam, and other services already give users, streamers, and creators tools to mute, block, kick, and ban troublemakers. However, tools exist to readily acquire new accounts and replace games at little or no cost -- things that easily circumvent even the most stringent cooperative-service ban suggested in this article. Most services make money off their users, so removing users is something that goes against their business interests too. The balance between keeping people safe and keeping the money flowing can very easily tip in either direction and fast.
I think the ultimate solution to the problem raised in this article is to put effort into understanding why people cause trouble. Gaming is a very big tent now. Gamers come from all backgrounds, ages, races, genders, skill levels, motivations for playing, expectations, and so on. Taking action against individual users hasn't worked long-term for decades. Also, what is toxic to one user is normal to another (e.g. trash talk is expected in competitive environments, but casual gamers consider it harassment). We should try separating people based on why they play games or what they're looking to accomplish. Use that machine learning and AI stuff to identify gamers that are likely to clash with each other, then keep them apart.
tl;dr -- The problem is real, but bans and policies haven't worked. Take a new approach like using AI to identify players that won't get along with each other and then keep them separated.