Snapchat及其競爭對手正努力應對人工智慧生成內容的混亂局面

Snapchat is taking steps to reduce the amount of fake, computer-generated videos on its platform. The company announced that it will stop recommending videos that are made entirely by artificial intelligence, or AI, in its Spotlight feed. This move is part of a larger effort by several popular platforms, including YouTube, LinkedIn, and Substack, to fight against the spread of fake writing, images, and videos created by AI tools. These tools have become easier to use, and as a result, more and more fake content is appearing online.

Snapchat's parent company, Snap, said that it wants to focus on showing 'authentic, human-made content' instead of videos that are completely made by AI. However, Snap is not banning all AI-generated content. The platform still allows users to use AI tools to edit or enhance their videos, and these edited videos can still be recommended to others. Snap explained that videos made only by AI are often low-quality, repetitive, and not what most Snapchat users want to see. Recent studies show that people agree with this view. When people see too much fake content, they start to doubt whether anything online is real.

Other platforms are also taking action. Substack, a newsletter website, has created a tool to help readers detect AI-generated writing. The CEO of Substack said that up to 40 percent of writing on social media is now fake or made by AI. He warned that platforms that reward fake content will create a 'race to the bottom. ' LinkedIn, a social media site for professionals, introduced a button that lets users report posts or comments that seem to be made by AI. LinkedIn's chief product officer said that fighting 'AI slop' is a top priority, and the platform has blocked billions of attempts to post AI-generated comments. LinkedIn is not banning the use of AI completely, but it is removing an automated prompt that offered to enhance posts with AI. Instead, it will go back to a simple proofreading tool.

YouTube, owned by Google, has also updated its rules about what kind of content can earn money. Research found that many YouTube channels were filled with fake, AI-generated videos, some of which had millions of subscribers and made millions of dollars. To discourage this, YouTube now has three categories of videos that cannot be monetized: generic, repetitive, or template-based. These types of videos are easy to make with AI tools. YouTube's trust and safety chief said that while AI can help people create good content, it is also being used to make fake and low-quality videos. YouTube wants to avoid 'content farming,' where people use AI to quickly make lots of bad videos.

All these changes show that big platforms are trying to make the internet a better place. They want to encourage real, human-made content and make it easier for users to trust what they see online. While AI can be useful for editing and improving content, these companies are making sure that fake, low-quality material does not take over their platforms. This is important for keeping the internet safe and trustworthy for everyone.