Baidu's Baijiahao Cracks Down on Self-Media Using AI to Spread Violative Content
-
Baidu's Baijiahao platform issued an announcement titled 'Governance on Combating Self-Media Using AI to Spread Violative Content,' stating that recent platform audits have revealed some authors using AIGC (Generative Artificial Intelligence) to create fake news and other violative content, severely impacting user experience.
In response to this phenomenon, the Baijiahao platform has implemented a series of measures, including taking down related violative content and imposing penalties such as account bans or suspensions on those involved in violative activities. As of November 30, 2023, the platform has removed 15,660 AI-generated violative content pieces and banned 489 related violative accounts.
Some typical behaviors of AIGC-generated fake and violative content include:
- Publishing unverified news without proper qualifications
Unverified accounts that use AIGC (Artificial Intelligence Generated Content) technology to mass-produce false news will be banned by the platform upon discovery.
- False Rumors:
Publishing false content related to public emergencies such as natural disasters, workplace safety, transportation incidents, or violent social events will result in the offending accounts being banned by the platform upon discovery.
Baidu stated that, in accordance with the "Notice on Strengthening the Management of 'Self-Media'," self-media accounts must accurately label the sources of information when publishing content related to domestic and international current affairs, public policies, or social events. For self-captured images or videos, each must be labeled with the time and location of capture. For AI-generated images or videos, they must be clearly marked as such. When referencing past events, the time and location of the original occurrence must be explicitly stated.
The platform encourages creators to declare AI-generated content. Creators can label their posts before publishing by navigating to "Publish Page - Settings - AI Content Declaration" and selecting the appropriate options.
The platform will continue to focus on addressing such issues, actively explore and upgrade governance methods and strategies, and regularly publish governance progress through platform announcements. We aim to comprehensively improve the quality and credibility of platform content, creating a healthy and civilized content ecosystem.