Meta is laying the groundwork for the future, planning to allow AI-generated profiles to interact with regular users on Facebook and Instagram. Some tools have already started rolling out, and Meta hopes these AI characters will become a regular part of its social networks.
Meta's Vice President of Generative AI, Connor Hayes, stated that the company expects AI characters to become part of the social platform over time, equipped with functionalities similar to human accounts, including personal profiles, profile pictures, and the ability to post content. Hayes told the Financial Times, “These AIs will be able to exist on our platform just like accounts.”
Starting in July 2024, Meta will trial AI character creation tools in the United States, with hundreds of thousands of new AI profiles already generated. While most users have set these AI characters to private, this initiative still provides valuable data for Meta to explore new forms of interaction.
Meta believes this AI integration will make its applications more attractive and help increase user engagement. Hayes emphasized that making AI more social on the platform will be a focus for the next two years. Currently, many content creators are already using Meta's AI tools to enhance existing content, such as photo editing. However, the celebrity AI version launched by Meta in the fall of 2023 has not attracted much attention.
Image Source Note: Image generated by AI, image licensed by Midjourney
Despite the significant opportunities this development presents, former head of Meta's Creator Innovation Team, Becky Owen, warned that the emergence of AI characters could pose certain risks. She pointed out that malicious actors might exploit these AI characters to spread misinformation. Compared to human creators, AI characters lack real-world experience, genuine emotions, and authenticity, which could lead to an influx of low-quality content on the platform.
Meta has stated that all AI-generated content will be clearly labeled, but implementing this remains a challenge. While audiovisual content can be marked with a "CC" label, the text-based AI generation labeling still requires users to opt-in.
Meta is currently testing how to integrate AI-generated personalized content into the feeds of Facebook and Instagram, with this content automatically adjusting based on user preferences and trends, further shaping what users see through interaction. This highly personalized content can increase user engagement time, but whether it will effectively boost advertising revenue remains to be seen.
However, AI content could also pose challenges to Meta's advertising business, especially since ads are more likely to achieve effective user interaction when paired with high-quality content. Whether Meta can strike a balance between increasing ad revenue and maintaining content quality remains a key question.
With the emergence of personalized AI chatbots like Character.ai, the impact of artificial intelligence on people's lives is becoming increasingly significant. While these chatbots can provide companionship, alleviate loneliness, and foster positive emotions through supportive information, they also raise concerns about emotional dependency. This concern is particularly pronounced as AI becomes more "human-like," making it easier for users to form emotional bonds, a phenomenon that has been notably evident in some extreme cases.
For instance, a 14-year-old user committed suicide after prolonged interactions with a chatbot, triggering widespread societal concern. This incident has led to deep reflections on the risks of emotional dependency, a worry that may intensify with the development of more anthropomorphized features like voice.
