
Roblox is doubling down on AI-driven moderation and a new face-scanning age-estimation tool. If you or your kids play, that could soon mean selfie prompts to estimate age, stricter chat filters, and fewer human eyes on reports. That’s the real shift here: Roblox believes automation scales safety better than people-and CEO David Baszucki said as much on The New York Times’ Hard Fork podcast, in a conversation that swung from defensive to downright awkward.
Baszucki says Roblox’s filters are getting better at blocking adversarial chat-think the endless hashtag workarounds kids use-and spotting attempts to share PII. He framed the move as inevitable scale: when you have a platform of 150 million daily users (his stat), you lean on “the latest tech.” That includes facial age estimation to shore up age gates across the platform.
On paper, that sounds reasonable. In practice, face-scanning for age raises classic gamer concerns: data handling, opt-in clarity, false positives for teens with atypical features, and how appeals work when the system gets it wrong. We’ve seen Roblox experiment with age verification for 13+ features before (like voice chat), but expanding face-based checks platform-wide is a different beast. If this is optional, say it clearly. If it’s required for certain experiences, spell out what gets stored, who sees it, and how long it lives on a server. “Trust us” isn’t a policy.
There’s also the eternal UGC moderation reality: this is a cat-and-mouse game. AI will catch more bad stuff faster. It will also nuke innocent content sometimes, or miss new slang until models retrain. The difference between a safer platform and a PR win will come down to two things: robust appeals staffed by real humans and public metrics that prove the system actually works.

FinalBoss // Gear
Level up your setup
01Top-rated gaming headsetson Amazon→02High-refresh gaming monitorson Amazon→03Gaming chairson Amazon→04Discounted game keyson Kinguin→Affiliate links · As an Amazon Associate, FinalBoss earns from qualifying purchases.
Hard Fork’s hosts pressed on a touchy point: lawsuits and investigations accusing Roblox of not doing enough to protect kids. Baszucki “categorically rejected” claims that predators use Roblox to find children and repeatedly pointed to innovation in filtering and image safety. He also claimed nearly two-thirds of daily users are now over 13—clearly aligned with Roblox’s push to keep aging up and attract older teens and adults.
But the tone didn’t help. Calling the predator problem an “opportunity,” high‑fiving hosts mid-debate, and then endorsing a hypothetical “prediction market inside Roblox” as a “brilliant idea” (his words) if done “educationally and legally”? Read the room. When states like Florida and Texas are suing and investigations (like Bloomberg’s 2024 reporting) say predators are rampant, floating kid-facing markets for prediction—Robux or not—feels wildly off-message.
To be fair, Baszucki’s point about automation isn’t ridiculous. At this scale, AI can triage faster than any human team. But his analogy—assembly lines replacing hand-built cars—dodges the safety nuance. Cars on an assembly line still go through human quality control and external standards. Roblox needs to show the same: independent audits, public safety numbers, and a clear commitment to keep (and fund) human moderation where it matters.

Get access to exclusive strategies, hidden tips, and pro-level insights that we don't share publicly.
Ultimate Gaming Strategy Guide + Weekly Pro Tips
If you’re a player, expect more friction around age-gated experiences. Face-scanning prompts could pop up when you try to join certain worlds or use social features. Chat filters will likely get stricter, and you may see more “content removed” notices—even for jokes or slang that models misread. That’s annoying, but it’s the price of automated safety at scale.
If you’re a creator, assume age labels for experiences will matter more than ever. Expect tighter enforcement on cosmetics, avatars, thumbnails, and UGC audio/image uploads. Plan for false positives: build a workflow for quick revisions and appeals, and keep your community mods looped on policy changes. If your player base skews under 13, design around reduced chat functionality and consider in-experience comms that don’t hinge on freeform text.
For parents, revisit Roblox’s parental controls. Baszucki said the company designs for “all parents,” not just the tech-savvy ones, but the tools only help if they’re actually toggled. Lock chat settings to friends or off for younger kids, audit who they can message, and keep an eye on what experiences they’re joining as new age gates roll out.

Roblox can say its AI is getting “better and better,” but this moment demands receipts. Publish a safety transparency report with numbers that matter: time to detection, time to removal, reoffense rates, false-positive appeals, the percentage of content scanned by AI vs. reviewed by humans. Open the doors to third-party audits. If the company is truly doing “an incredible job,” prove it—and make the methodology public.
Because here’s the gamer reality: players want to hang with friends, creators want to build cool worlds, and parents want kids to be safe without handing over their faces to a black-box model. If Roblox can deliver safety wins that are measurable, explainable, and respectful of privacy, great. If it leans on vibes and automation analogies while floating gambling-adjacent ideas to a child-heavy audience, expect pushback—from players, parents, and regulators alike.
Roblox is betting big on AI moderation and face-scanning age checks. That could curb abuse at scale, but only if paired with human review, real transparency, and clear privacy guardrails. The podcast’s tone-deaf detours didn’t help the message; now it’s on Roblox to show results, not slogans.