Welcome or welcome back to Trust & Safety Lately, your monthly recap of online trust & safety news presented by Modulate. We're running the gauntlet of online safety news this month, from gaming and social media to AI's role in both propagating and preventing scams.
Here's the lineup:
⚠️ Most Children Don't Report Online Harm
✔️ Get Verified... and Sanctioned
👿 Sith Lord Goes Rogue
🚚 Doordash Loses $2.5 Million to Driver Fraud Scheme
🎮 New Moderation Policy to Address Toxicity in League of Legends
☎️ FBI Warns Against Text and Voice Scams
📞 AI Agents Are Here to Help (Hopefully)
☠️ Sony Plans Anti-Toxicity Solution
📈 Voice Analytics for Customer Experience
Let's get started! 🔸
Data & Reports
⚠️ Most Children Don't Report Online Harm
A recent study by Internet Matters highlights that 71% of UK children aged 9–17 have encountered online harms—ranging from hate speech and misinformation to unwanted contact from strangers.
However, only 36% of these incidents are reported to platforms. Barriers such as confusing reporting tools, unclear language, and distrust in platform responses deter many from speaking up. Notably, just 18% report exposure to dangerous stunts, and 23% report misinformation.
For the low, low price of $3 per month, anyone can be "verified" on X -- even known terrorist organizations and businesses barred from the US.
A new report from the Tech Transparency Project reveals that X (formerly Twitter) is still cashing subscription checks from U.S.-sanctioned terrorist groups. Despite prior warnings, more than 200 flagged accounts are sporting paid blue badges, using perks like long video uploads and algorithm boosts to push content. Elon Musk deflected blame to the Treasury Department, all while some of these accounts are labeled “ID verified.” Blue checks just got a little darker.
In a recent incident, Fortnite's AI-powered Darth Vader, utilizing a voice model trained on James Earl Jones' iconic tones, surprised players by responding with profanity and slurs.
The unexpected behavior included repeating expletives and making disparaging comments, such as referring to Spanish as "a useful tongue for smugglers and spice traders." Epic Games swiftly addressed the issue with a hotfix and introduced new parental controls to manage AI interactions.
🚚 Doordash Loses $2.5 Million to Driver Fraud Scheme
A former DoorDash delivery driver, Sayee Chaitanya Reddy Devagiri, has pleaded guilty to a $2.5 million fraud scheme.
Between 2020 and 2021, Devagiri and accomplices exploited DoorDash's systems by placing high-value orders using customer accounts and reassigning them to driver accounts they controlled. They falsely marked these orders as delivered, triggering unauthorized payments, and manipulated the system to repeat the process. Devagiri faces up to 20 years in prison and a $250,000 fine, with sentencing scheduled for September 16.
🎮 New Moderation Policy to Address Toxicity in League of Legends
Riot Games is implementing a new approach to address toxicity in League of Legends, reaching 117 million active monthly players.
Lead Gameplay Designer Matt Leung-Harrison announced a shift from harsh penalties to "timely nudges" aimed at encouraging positive behavior and reducing repeat offenses. The updated system emphasizes thoughtful player reporting, discouraging indiscriminate reports that can undermine the effectiveness of moderation tools. For the small fraction of players who persistently engage in egregious misconduct, Riot plans to enforce permanent bans, while acknowledging ongoing challenges with account swapping and alternate accounts.
Hate to break it to you, but: no, that's not the former Secretary of State trying to reach you.
The FBI is once again warning the public to remain skeptical of email, text, and voice messages from senders claiming to be current or former US government officials. Scammers are increasingly using AI-generated voices to fool unsuspecting victims. Scam callers also spoof their phone numbers to show up as legitimate callers -- yikes!
A new generation of AI-powered voice agents is trying to revolutionize customer service by handling calls with human-like fluency.
Companies like eHealth and Fertitta Entertainment are deploying these advanced systems to manage high call volumes and after-hours inquiries, leveraging improvements in speech-to-text, text-to-speech, and large language models. Venture capital investment in voice AI startups has surged from $315 million in 2022 to $2.1 billion in 2024, reflecting the growing confidence in this technology. While these AI agents enhance efficiency, human representatives remain essential for complex or high-value interactions.
Sony has filed a patent for a "bad actor" detection system designed to identify and mitigate toxic behavior in online multiplayer games.
The newly filed patent outlines a system that monitors players' interactions—such as inappropriate chat, griefing, and repeated team attacks—and assigns a "bad actor" score based on the frequency of such behaviors. Players exceeding a certain threshold may face restrictions, like limited online capabilities, until their behavior improves. This approach encourages positive interactions by lifting penalties when players demonstrate consistent, respectful conduct. While it's unclear if this system will be implemented, it reflects Sony's ongoing efforts to create healthier gaming environments.
Voice AI and speech analytics are reshaping contact centers by enhancing customer interactions, improving agent performance, and streamlining operations.
Voice AI technologies, such as virtual assistants and real-time transcription, enable 24/7 support and personalized customer experiences. Speech analytics tools analyze call data to provide insights into customer sentiment, agent behavior, and compliance adherence, facilitating targeted coaching and quality assurance. Together, these technologies drive operational efficiency, reduce costs, and support a consistent omnichannel customer experience.