All articles tagged with #ai moderation
YouTube's AI moderation system has been removing videos that show how to bypass Windows 11 restrictions, leading to concerns about overreach and lack of human oversight, with some speculating possible pressure from Microsoft. Creators are frustrated by the automated process and lack of transparency, fearing it may stifle technical content and free expression.
Many Instagram users worldwide are experiencing wrongful account bans due to Meta's AI moderation, leading to emotional distress, loss of personal and business content, and concerns over police involvement. Meta has acknowledged some issues but denies broader problems, often reinstating accounts after user complaints, though frustrations remain over the process and accuracy of automated enforcement.
The article exposes the hidden human toll of AI content moderation, highlighting how workers, mainly in the Global South, face psychological trauma, exploitation, and silence due to NDAs and outsourcing practices. It calls for global reforms to protect these workers' rights, mental health, and to dismantle the oppressive systems that profit from their suffering.
Call of Duty's new AI-powered voice chat moderation tool has taken enforcement action against over 2 million players for toxic voice chat, resulting in global muting and restrictions for repeat offenders. The system has led to an 8% reduction in repeat offenders and a 50% reduction in severe instances of disruptive voice chat since its launch. Activision is encouraging more user reports and is committed to evolving the moderation technology to combat toxic behavior in both voice and text chat within the game.
Activision has partnered with Modulate to introduce an AI moderation system called ToxMod in Call of Duty games, aiming to identify and address hate speech, discrimination, and harassment in real time during voice chats. The beta rollout of ToxMod has begun in North America, and a worldwide release is scheduled for November 10th. While the AI tool will not take direct action against players, it will provide reports to Activision's moderators. Human involvement remains crucial to address potential biases in speech recognition systems.
Microsoft has launched Azure AI Content Safety, an AI-powered moderation service that can detect inappropriate content across images and text in multiple languages. The service assigns a severity score to flagged content, indicating to moderators what content requires action. Azure AI Content Safety is integrated into Azure OpenAI Service and can be applied to non-AI systems, such as online communities and gaming platforms. Pricing starts at $1.50 per 1,000 images and $0.75 per 1,000 text records.