Dubai: A dialogue session at the World Police Summit 2026 on protecting children through the Roblox electronic gaming platform witnessed a confrontation between a number of parents and the head of the company’s law enforcement communications department for the Middle East, North Africa, Turkey, the Caucasus and Central Asia region, Rafik Hounous.
Their concerns regarded protection and security factors on the platform, and the possibilities of children being exposed to cybercrimes and threats, including grooming and blackmail by adults who may manipulate their ages and identities to reach them.
These discussions come at a time when Roblox is facing increasing legal and regulatory pressure in the United States, on the back of accusations for not providing sufficient protection for children from sexual exploitation, grooming attempts and inappropriate content, along with claims about inadequate age verification tools and parental controls, and presenting a misleading picture to families about the level of safety on the platform.
During the session, Hounous said that protecting children is a fundamental responsibility of the company, noting that the platform adopts a set of technical and monitoring procedures, including estimating users’ ages through facial recognition, classifying accounts according to age groups, and reviewing content and conversations using artificial intelligence and human reviewers, in addition to cooperating with law enforcement agencies in cases involving threats or exploitation of children.
He pointed out that a child sent a message revealing that he had been assaulted by his father in one of the countries, and it was dealt with seriously, and it was discovered that he was not the only son who was subjected to this by the father.
Parents' comments focused on the extent to which these measures can counter the methods that criminals may resort to through the platform, especially attempts to hide identity, use accounts with fake ages, and exploit symbols, innuendos, and indirect phrases to lure children away from family supervision and the platform's systems.
Concerns raised by one mother during the session included the possibility that children could circumvent the age verification system by using the face of an older sibling or family member to bypass facial recognition technology and access content or features not intended for their age group.
In response, Hounous said the company conducts frequent user identity checks, likening the procedures to customer identity verification mechanisms in banks, as part of Know Your Customer (KYC) and anti-money laundering and counter-terrorism financing requirements, with the aim of ensuring that the person using the account is indeed its owner.
He explained that the facial age estimation technology has, according to the company's estimate, a margin of error of about one year, and that it advises parents to repeat the verification process as the child gets older, so that he can move to the appropriate age group.
He stated that the platform adopts three main categories of users: the first for children aged five to eight years, the second for the category of nine to 15 years, in addition to the general category for those aged 16 years and older, with different levels of content, conversation features and parental control tools between each category.
The attendees' concerns extended to the ability of artificial intelligence systems to detect attempts to lure children that do not rely on direct words, but rather use hints, emojis, innuendos, and coded phrases, allowing harassers to bypass traditional monitoring systems.
In his response, Hounous acknowledged that there is no perfect artificial intelligence system, explaining that the company may disable some features in countries where there are not sufficiently effective tools for reviewing content and conversations, until more suitable technologies are developed.
He added that the company is working to improve monitoring capabilities in multiple languages and dialects, relying on a combination of artificial intelligence and human review, noting that some countries do not have conversational features due to insufficient review tools.
Parents were urged not to rely solely on the platform's tools, but to monitor the games their children use, talk to them about the people they communicate with, activate parental control tools, and educate them on how to deal with suspicious messages or behaviours.
Arabic-language restrictions
He revealed that the platform sees an exchange of about 6.5 million messages daily, noting that the company's systems, according to his estimate, can detect personally identifiable information about users in English-language conversations up to 98%, with this information being blocked as soon as it is detected.
In contrast, he acknowledged that the company currently lacks an effective level of review for conversations in Arabic, which prompted it to completely disable the chat feature in the UAE, regardless of the user's age.
He explained that the conversations on the platform are not encrypted, which allows safety and security teams to view the content when suspicious activity is detected, and to assist the competent authorities in dealing with reports and investigations.
He said the company is seeking to make the protection settings available by default, in response to feedback from parents in the UAE and the Gulf countries, while allowing parents to activate some features if they wish.
He added that parents who link their accounts to their children's accounts can receive weekly emails that include a summary of the child's activity, including the games they used, the number of hours used, and the people they communicated with.
Regarding dealing with cases of exploitation and threats, Hounous confirmed the existence of proactive cooperation with law enforcement agencies, noting that the company refers cases of child exploitation to the US National Center for Missing and Exploited Children, which shares information with partner agencies, and also uses channels linked to the US Federal Bureau of Investigation in some cases related to security threats.
He added that the company is seeking to develop more specific communication channels with law enforcement agencies, allowing for proactive reporting of potential threats, such as school shootings, online child grooming, and putting children at risk.
He cited an incident that he said occurred several months ago, when the company's systems detected a message from a child in one of the countries stating that he was being abused by his father. The company filed a report with the competent police, who went to the house and discovered, according to his account, that the child and two other children, along with the wife, had been abused.