Mark Zuckerberg, the CEO of Meta (formerly Facebook), recently shared his plans to integrate generative artificial intelligence (AI) into popular platforms like Facebook, Instagram, and Messenger.
Generative AI has gained popularity with apps like Dall-E, ChatGPT, and Midjourney, allowing users to create various content. Now, major tech companies like Google, Microsoft, and Adobe are also incorporating generative AI into their products.
During an internal meeting, Zuckerberg revealed Meta's intention to include text, image, and video generators powered by generative AI in Instagram, Facebook, and Messenger. He stressed the significant breakthroughs in generative AI over the past year and expressed the company's determination to incorporate this technology into all of their products.
Meta has been investing heavily in AI research and shifted its focus to metaverse technology in late 2021. In February, the company restructured its AI team to facilitate the integration of generative AI into their products.
Specific examples of how Meta plans to use generative AI were discussed during the meeting. One example is a text prompt feature that allows Instagram users to modify their photos before sharing them. Another possibility is the addition of AI-powered chatbots, similar to Snapchat's, to Messenger and WhatsApp, offering a range of personalities and capabilities.
Although the timeline for implementing these features remains unclear, Meta aims to make generative AI a core part of its flagship products. The company is actively working on incorporating generative AI into the metaverse, with a focus on creating 3D visuals.
To encourage innovation, Meta is granting its employees access to internal generative AI tools and organising a hackathon to showcase AI projects. Additionally, they plan to introduce a service on Instagram that enables users to modify photos using text prompts and share them in Stories.
While Meta is committed to sharing AI research with the open-source community, they have not addressed recent concerns about the leak of their LLaMA language model.