
After parking Lada Granta car in parking lot in the vicinity of Luzhniki Stadium, located
in Moscow, when Ivan clad in jacket and Cossack hat disembarked to see the soccer match
between Russia and England, his Russian friends друг encircled him, hugged and chat. The referee whistled and the fantastic match kicked off.
To support Russian players, famous Russian celebrities were also present in the stadium to
watch with enthusiasm among millions of crowd and fans. While drinking aromatic cup
of Chernaya Karta coffee, Ivan got alert message in his smartphone so when he opened his
Gmail inbox, found tech article and headline was Odnoklassniki (OK) now has friendly AI that mitigates toxicity.
The social network’s creators of OK have announced the release of a machine learning-based
AI model that restricts invasive dating and communication with uninvited parties.
In order to ensure that victims of cyberbullying never confront their attackers, it covertly examines users.
OK social network
By eliminating needless notifications from obtrusive interlocutors, the new algorithm hopes
to alleviate users who frequently decline to communicate and stop “stalkers” from upsetting others.
The social network OK has figured out how to tell one from the other in order to accomplish this.
The algorithm examines a user’s social interactions, including offers and requests for
friendship, page visits, gifts, and a variety of other factors, to identify the sort of person.
When someone consistently adds new friends without knowing them, delivers gifts to
random individuals in bulk, assigns them “classes,” and sends them private messages, the neural network classifies them as obsessive.
Individuals who frequently decline friend requests, ban users, turn down presents, or
voice grievances about strangers’ behavior are include in the AI model.
Recent data indicates that 42% of users polled think that over the past few years, there has been
a noticeable increase in the incidence of cyberbullying.
The social network OK hides alerts from prying eyes and restricts communication between
these two types of accounts based on the information gathered.
The second group of users, for instance, will never notice a friend request, gift, or page visit from a “stalker.”
Additionally, fewer talks between the aforementioned groups possible by the
neural network; the quantity of private messages sent by bulk senders will be constrain.
During several months, the neural network tested on several social network audiences. As part of the trial, there were 25% fewer complaints about unfamiliar accounts causing invasive attention than in earlier periods.
The tendency for users who had previously avoided communication to reject friend invitations
and put people on their “black list” decreased as unpleasant activities stopped drawing notice
and the friend requests themselves gained significance.
When finished reading tech news, Ivan shared it on OK social network.
