Aiba - Safe Digital Lives

A once-active player goes quiet. The messages get shorter. Eventually, they stop showing up. That doesn’t automatically mean harm occurred. But it raises a question: What might we miss when reports are our main measure of safety? This article explores the patterns worth noticing, and why context and human judgment matter. https://lnkd.in/eDnekRpd 🫡 💚 🚀

Debate: How do we build trust and safety in children's digital lives? Age limits. Bans. Education. Everyone has an opinion on what works, few have real answers. This week, Aiba's Hege Tokerud joins a public debate in Gjøvik, Norway alongside social commentator Hanne-Lene Dahlgren, Professor Patrick Bours (NTNU Gjøvik) and deputy mayor Kjersti Bjørnstad to talk about online risk and what parents and young people can actually do about it, starting today. Professor Patrick Bours's research on behavioral biometrics is what led to Aiba's founding, so this one is close to home. The event is in Norwegian, but the question it's asking is one every platform and every parent is wrestling with right now. Link in the comments

The EU just proposed a law that treats video games the same as social media and AI chatbots. The EU KIDS Act, proposed 17 September, flips the burden of proof: platforms will have to show their service is safe for minors before problems appear. The proposal has 4 pillars: 👉 A social media age ladder: no accounts under 13, guardian managed accounts from 13, full accounts from 15 👉 Safety by design: no addictive feeds, no unsolicited contact, AI companions off by default for minors 👉 Privacy preserving age checks 👉Enforcement built on the same structure as the DSA and AI Act, with compliance plans and audits for the largest platforms Keep in mind this is still a proposal, but the direction is clear. We had a closer look at what this might mean for platforms that minors use. Link in the comments

The internet could use a “nabokjerring.” In Norway, that’s the neighbour who notices what’s happening. Who steps in when someone is being cruel, and checks on the person who got hurt. Tough and kind in equal measure. Online communities need that kind of care too. AI can help moderation teams spot harmful patterns earlier. But someone still needs to understand the situation, make a fair decision, and respond to the person who needs help. A report deserves a proper look. A moderator needs support with a difficult call. A player needs to know someone is listening. These everyday moments shape whether people feel safe enough to stay, participate and trust a community. At Aiba, this is the work we want technology to support. We’ve written about the people behind safer communities, the limits of automation, and why looking out for one another remains everyone’s responsibility. Read the article: https://lnkd.in/e7_xyMks 💚 🚀 #TrustAndSafety #OnlineSafety #ContentModeration #CommunityManagement #Aiba

Would you know what to look for if a child you care about was experiencing online grooming? For parents and caregivers, knowing which signs to look for can be a starting point for an important conversation. In our latest article, we highlight signs that may indicate a child is experiencing online grooming. Read and share the article below, and help us reach more parents and caregivers by sharing it with your network. https://lnkd.in/eTBRiz8S 🚀

Most studios approach moderation the same way: Block the bad words and deal with the harm when someone has reported it, very often after it has happened. A filter, a word list, an auto-flag when someone types something ugly. Job done, community safe. Except the most dangerous conversations in your game contain no bad words at all. Grooming starts as friendly attention. Manipulation starts with isolation, a kid being told nobody else cares about them. By the time the language turns explicit, the harm is already moving off your platform. And when a real flag finally lands on a moderator's desk, the decision is difficult. Ban this user. Report this to the police. Or sometimes: this is a child showing signs of self-harm, often the result of bullying, and someone needs to respond like a human being. Underneath all of this sits an uncomfortable trade-off. You can tune a detection model to produce fewer false positives. Less noise, less manual review, a cleaner dashboard. But every conversation you filter out of the queue is one nobody will ever look at. And some of them are real. At Aiba, we made that trade-off the customer's choice. Every studio decides how many false positives they're willing to review. Every single one has chosen the more sensitive setting. More flags, more manual review, on purpose. They'd rather check a false alarm than miss one real case of harm. On discovery calls I often hear the opposite instinct: "the AI should just take care of it, it should be automated." I understand where it comes from. But the flag is not the decision. We believe in a human on top of the system, not just in the loop. Our customers get it. Their community is the heart of their game, and they take care of their players accordingly. Grooming is one of the hardest problems on the internet. It hides in plain text, and it's easy to miss. It's happening right now, and it can be solved. Not with a word list. Not with full automation. With technology that finds the pattern early, and people who decide what happens next. Bad actors find new ways in every day. So we improve every day. That's the job. Aiba - Safe Digital Lives is the firewall for humans.

The moment a predator moves a kid off your platform, it's game over. And getting them off a platform can happen in a few friendly messages. The primary way grooming happens is not on your game. It starts there. Then someone takes the conversation to Snapchat, Discord, or WhatsApp, where nobody is watching and you can't see a thing. These people even have a name: harvesters. They're after one thing, a username. And getting one can take seconds. Now, I know the pushback. People meet in games, build real friendships, even get married. True. Those are great stories, and we have to make sure they keep happening. Most people are nice. It's the bad apples with bad intentions causing the harm. But if we assume everyone online has good intentions and stop paying attention, it's always the vulnerable who pay the cost of that way of thinking. Some kids are struggling. At home, at school, they're left out. Then they meet someone who sees them, and they feel seen. That's exactly the kid a predator is looking for. This is why detection speed is everything to us. In 2021, we could detect grooming in under 40 messages. Nobody had ever done that. Today we do it in under 10, at the very start of a conversation, right when the "what's your Snapchat" questions begin. Every message earlier means we catch it while the child is still on your platform. Still visible. Still protectable. But detection is only half the picture. The other half is your own community. Start with your moderation guidelines. Know what kind of game you run, kids, teens, adults, because each has different needs. And don't assume this is a kids'-game problem: in adult games we see social manipulation too. Adults get tricked into sending pictures, then blackmailed. It has happened to people I know. Context is everything. That's what we're actually built to find: the context and the behavior, together. Happy to chat if you run any kind of online community.

A turning point for child safety online? The Meta settlement signals an important shift: child safety is no longer only about removing harmful content after the fact. It is increasingly about how platforms are designed, governed, and held accountable. The real test will be whether these safeguards become a lasting industry standard rather than an exception. Ben Whitelaw shares five thoughtful takeaways on what the settlement could mean for platforms, regulators, and the future of online safety: https://lnkd.in/dPccyPVv #TrustAndSafety #OnlineSafety #ChildSafety #ContentModeration #SafetyByDesign