This year at Roblox RDC25 (Roblox Developer Conference 2025), the Builderman Q&A (Question and Answer) along with the new Safety Q&A (Question and Answer) was private and kept to attendees only. Roblox has recently released an anonymized transcript of the questions and answers.

False positives and creator bans

Nick admitted false positives happen (like harmless assets being flagged) but said Roblox takes them very seriously. The company is working on reputation-based moderation — long-standing good creators won’t be punished the same way as repeat bad actors. Punishments will also scale by severity, so small mistakes don’t get the same response as egregious violations.

Preventing unsafe AI-generated content

Matt explained Roblox already moderates its own GenAI models with prompt filtering and post-generation checks. They even open-sourced some tools so others can use them. For creators using external AI, Roblox is building a moderation API that can check models, images, and audio in real time. Eliza emphasized that developers won’t be blamed for bad user-generated content — accountability will fall on the actual bad actors.

Responding to “virtue signaling” concerns

A questioner asked how Roblox can prove its safety work is genuine, not just PR. Matt pointed to age verification for all communication features as one real step. Eliza noted most safety work is invisible (“if we do our job perfectly, nothing happens”), but Roblox will start communicating more openly — like open-sourcing models, sharing law enforcement partnerships, and showcasing proactive measures. Nick added Roblox holds itself to higher standards than any other platform and sees it as a chance to lead the industry.

Parental involvement and safe defaults

Eliza outlined recent upgrades to parental tools: screen time insights, spend controls, and connection oversight. But instead of relying on parents to tweak settings, Roblox wants to make the platform “safe by default” based on verified age. She also revealed Roblox has hired a Head of Parental Advocacy and is launching a Teen Council to boost digital literacy for both teens and parents.

The Discord problem: off-platform behavior

A developer raised concerns about harmful behavior on Discord spilling over into Roblox communities. Matt acknowledged the difficulty: while platforms do cooperate on illegal threats, Roblox can’t always act on behavior that isn’t a violation elsewhere. Privacy rules also make it hard to verify evidence across platforms. Roblox does try to validate external evidence before acting, and discussions with other platforms are ongoing, but Matt stressed: “It’s not a simple solution. I wish it was.”

Privacy and age verification

A creator raised concerns about uploading government IDs or selfies to verify age. Matt acknowledged the tension: Roblox has to confirm age for safety but also values privacy deeply. Current scalable options are limited — ID scan + face match, or facial age estimation. Roblox uses partners today but is exploring safer, lower-risk methods (and hopes Google, Apple, or Microsoft will eventually provide built-in solutions). Privacy, he stressed, is “built into our DNA” as Roblox openly designs for kids.

Transparency tools for advocates

A long-time developer asked for better ways to counter viral “bad Roblox” narratives. Eliza pointed to the transparency reporting site (showing takedowns, law enforcement referrals, NCMEC reporting, etc.) and Roblox’s newsroom posts about proactive safety launches like age estimation and parental controls. Nick asked what would help more, and the creator said the info often feels too corporate to use in conversation. Eliza suggested the Community Safety Council could co-create digestible “one-pagers” or shareable stats. Matt added that Roblox is the only platform monitoring all user communication and regularly reports thousands of serious cases to law enforcement — steps that set it apart, even if not always visible.

Shifts in safety strategy amid scale and scrutiny

A father/dev asked how Roblox’s safety approach has shifted with recent media and political attention. Matt said Roblox used to avoid broad sweeps of old content to prevent unfairly catching innocent creators. But with Roblox’s cultural visibility, “false equivalencies” (like usernames mistaken as signals of systemic problems) can’t be ignored anymore. Roblox is now being more aggressive in cleaning up legacy content and proactively engaging with governments worldwide as regulation tightens. Recent policy changes in the Middle East were cited as one outcome of this shift.

Enforcing rules when games have bad actors

A dev asked about Section 6 of the Creator Terms, which lets Roblox moderate experiences if too many users violate rules — raising MeepCity as an example. Eliza said the clause exists specifically to address cases where the creator isn’t at fault but the community behavior is out of control. In the past, moderation teams manually counted violations in games, which wasn’t scalable. Now, new “bad scenes” detection technology can flag when a game tips into “mostly bad behavior.” When Roblox contacts devs, she said, they’ve always responded quickly with fixes. The goal is to eventually give devs a dashboard with early warning signals before formal enforcement. Eliza also cautioned about abuse vectors (like griefers trying to get a rival game flagged), so Roblox is designing guardrails. Lauren added that Roblox has been inspired by developer creativity when solving these issues — some solutions from one game have been recommended to others.

Privacy and ID verification concerns

A developer raised worries about having to upload government IDs or selfies online. Matt acknowledged the concern and said Roblox builds privacy into its systems since it’s one of the only platforms openly acknowledging children on it. Current scalable options for age verification are ID + face match or facial age estimation. Roblox uses a partner for this today but is actively exploring safer and more privacy-preserving solutions, including potential integrations with Apple, Google, or Microsoft.

Advocating for Roblox in public conversations

A community member noted that it’s hard to counter viral negative content about Roblox because official transparency often feels “corporate-y.” Eliza pointed to Roblox’s transparency reports (including takedowns, law enforcement referrals, and NCMEC reports) and promised more proactive storytelling. Nick asked what would help, and the questioner suggested simpler, shareable talking points. Eliza agreed and floated the idea of one-pagers and stats for the community to use. Matt added that Roblox is the only platform monitoring all user communications and reporting thousands of cases to law enforcement each year — but acknowledged the company needs to better explain this work.

Shifts in approach after recent press scrutiny

A father and developer asked how Roblox’s safety thinking has changed amid heightened media attention. Matt said Roblox used to avoid sweeping “back catalog” moderation to prevent overreach, but false equivalency headlines (like usernames being tied to broader safety problems) have forced a more aggressive cleanup strategy. Roblox is also engaging governments worldwide more proactively as regulation of online spaces increases.

Responsibility for safety in games with bad actors

A developer asked how Roblox applies the policy that allows moderation of experiences with widespread rule-breaking users (like MeepCity). Eliza said the rule exists to address bad user behavior without blaming developers. Historically, moderation relied on staff manually checking games, but new “bad scenes detection” tech lets Roblox measure when a game tips into egregious behavior. Developers contacted by Roblox have been very responsive and creative in fixing issues, and Roblox hopes to eventually provide early warning dashboards. Guardrails are also needed to prevent abuse (like griefers trying to get games taken down).

Advertising and age-appropriate standards

Asked about ads being safe by default, Eliza said no ads are shown to under-13 users, and strict ad category rules are enforced. All ads must follow Roblox’s community standards. Matt added that Roblox moderates ad partners, serves contextual ads under 18, and restricts rewarded video ads for under-13s. Policies are continuously updated to match regulations worldwide.

Inappropriate avatar combinations and possible age-gating

A questioner noted that accessories are often fine individually but inappropriate in combination. Matt admitted this is a huge challenge at scale — billions of avatar combinations occur daily. Roblox is working on scalable solutions. Eliza added Roblox is considering a maturity rating system for avatar items, similar to games, so some accessories may eventually be age-gated.

Age verification for future “dating” and 18+ content

A developer asked if ID verification will be required for future dating features. Matt said Roblox’s vision is to support more aged-up, 18+ content (like R-rated style games) while ensuring exposure is age-appropriate. Current scalable solutions are ID verification and facial age estimation, with openness to new tech in the future. Eliza added that while ID is more clunky and inequitable (since not everyone has one), facial age estimation can be more inclusive if accurate enough. For sensitive use cases like dating, Roblox would only rely on tech with proven reliability.

About the Author

ScrapBlox

Administrator

Owner of Robloxia.org | Need help, want to become an editor? [email protected]

View All Posts