TikTok is introducing a new feature that enables them to reduce the quantity of artificial intelligence-produced media appearing in their streams. This initiative comes as the service announced that it presently contains more than 1 billion AI-generated clips.
Testing and Rollout
The new functionality is in a pilot phase over the coming weeks before a widespread rollout. This shift comes amid a surge in AI content, driven by recent video-generation tools.
AI Content on Other Platforms
It has been noted that nearly 10% of the most rapidly expanding video streams on a competing service only show content made by AI. A large amount of this content is classified as mass-produced AI material, a label for substandard, mass-produced content that is typically absurd or bizarre.
Our research shows that a lot of users enjoy content produced via AI tools, from computer-generated images to science explainers, and we aim to provide users the power to increase or decrease of that, according to their own likes,” stated a company executive.
Scale of AI Content
TikTok shared that there are currently 1.3bn videos on its app labeled as artificially created. Overall, more than 100 mn pieces of content are uploaded to the app each day, making AI-generated material a relatively small part of its complete catalog.
User Control Features
Users will be able to decrease or raise the level of AI content they watch by opening the “manage topic” option in the app and picking AI-made media. Further subjects that are already be controlled under the setting cover current affairs, clothing, cosmetics, and dance.
Content Labeling and Policies
The platform’s rules require content makers to identify convincing artificially created content, and it forbids malicious AI impersonations of celebrities or critical incidents. Any authentic-looking artificial content that is unmarked can be removed under its community policies.
In addition, the service will now apply an AI-generated label to content made using its proprietary AI systems, or identified by an cross-platform project for recognizing content made by AI. This step is designed to deter users from trying to bypass the labeling process.
Responsible AI Funding
The platform is also establishing a $2m educational fund for specialists and institutions to create informative material about using AI responsibly. This initiative will support charitable organizations focused on educating computer science.
Content Safety Approach
The platform has supported its moderation strategy, amid discussion over intentions to make redundant hundreds of safety staff in the United Kingdom. The company is planning to eliminate hundreds of positions in its trust and safety team in the UK capital, with worker representatives and digital protection specialists raising concerns that staff are being displaced by automated systems.
A trust and safety lead stated that people-led review would remain essential, but artificial intelligence can safeguard staff by filtering out the most distressing content before it is viewed by staff. The company reported a 76% decrease in explicit and harmful content reviewed by safety staff over the past year due to reliance on its AI tools.
We believe it is crucial to combine humans and technology to ensure user safety,” she said. “We’ll have people as involved in the system, but we also think it’s very important to minimize viewing of unsafe media as swiftly, which is where a lot of the machine support is making a difference operate efficiently.”
