Meta has officially entered the race of chatbots and conversational AI with the launch of Meta AI and numerous other AI bots. In what has been a rather major launch, Meta announced a slew of chatbots for its social media platforms like Messenger, WhatsApp, Instagram, and more. Meta AI and the other AI offerings are based on generative AI technology derived from language models, given that the firm spent a considerable amount of time developing and refining its own LLM—LlaMa 2. Meta AI will also be available on Meta’s wearable offerings like the Ray-Ban Meta smart glasses and the firm’s VR offering titled “Quest 3”. In addition to communicating with users, Meta AI can also generate realistic images that can be shared with other users. The image generation protocol is named Emu and was trained with over 1.1 million image and text pairs, which also included images and captions shared on Facebook and Instagram.
Meta has partnered with Microsoft’s Bing to bring about an internet search feature within its chatbot interface. This makes Meta AI up-to-date in its information, allowing users to access the internet from within the conversational AI chatbot. Alongside these features, Meta has also launched “Gen AI Personas,” which are essentially AI chatbots with defined precepts that dictate their conversational tone, style, and behavior. Imbued with a personality based on popular influencers or celebrities, Meta is positioning this product specifically to retain interest from the younger generation. As other competitors like OpenAI and Google explore numerous means to target a broader client base through plugins and expansions to existing frameworks, Meta seems to have taken a different approach.
An Overview of Meta AI, Gen AI Personas, and Meta’s Other AI Offerings
Meta AI will be available on the firm’s social media platforms.
Meta unveiled 28 other chatbots in addition to Meta AI, essentially bringing forth a trove of several offerings for users to interact with. All of these models are based on customized versions of the LlaMa 2 language model which Meta developed in collaboration with Microsoft. Meta AI boasts several interesting features and capabilities that make it a versatile offering in the growing list of AI chatbots and LLMs. The chatbot is designed to be interacted with like one would with a person, much like what Snapchat did with My AI. In Meta AI’s case, the firm also launched an AI Studio for businesses to design and curate their own conversational AI chatbots to interact with prospective customers and clients. Interestingly, Meta AI is not restricted to merely an individual chat window of its own and can also be called upon in conversations with other humans. However, the chatbot will have to be specifically mentioned within the chat to evoke a response. Though this might throw up privacy and safety concerns, the chatbot only reads and responds to messages it’s tagged in.
While Mark Zuckerberg admits that chatbots are still a work in progress, he mentions Meta’s approach toward AI and conversational AI chatbots leans more toward entertainment rather than other serious use cases. The inspiration for the creation of Gen AI Personas comes from the intent to galvanize engagement with younger user groups. Meta has based these chatbots on popular personalities including famed rapper Snoop Dogg and sports stars like Tom Brady. Numerous other celebrities, too, have signed up to lend their personalities to Meta’s chatbot and have partnered with the firm to model an AI interface that emulates their behavior. Similar machine learning protocols have been utilized and deployed by other AI firms such as Character.ai. Apart from these intuitive chatbots, Meta has also added AI-based editing features for users’ images. In addition to its in-house image generation protocol, Meta AI might just become an important competitor to Dall-E and Midjourney.
Meta AI’s Technical Attributes and Endowments: Is the New Chatbot Series a ChatGPT Competitor?
Meta AI includes numerous functionalities within its ambit.
Meta AI’s framework is based on an advanced architecture derived from Meta’s LlaMa 2 language model, which was released earlier in 2023. The chatbot has an ingenious ability to detect user intent through an in-built orchestrator that functions on complex natural language processing algorithms. In addition to driving engagement, Meta has also positioned Meta AI as an alternative to other prevailing chatbots in the markets such as ChatGPT and Bard. While these competitor chatbots might be used and deployed in a more formal use case, Meta seems to have caught up to the rivalries in the industry. The company also seems to have deployed advanced hardware and processing capabilities in its development of Meta AI, having used frameworks with computer vision capable of generating information on the environment by merely analyzing videos and images.
To add to existing generative capabilities, Meta AI has also been endowed with unsupervised machine translation, further closing the gap between languages and their understanding. Safety of communication has also been an area of central focus, with Meta having included these features to ensure resilience to security threats regardless of the language being used. The current frameworks in place elicit Meta’s broader AI vision and investments, though Meta AI might be limited to entertainment and business purposes. Alongside advanced technical capabilities, Meta also ensures all user data is used responsibly. As firms like Meta further their AI capabilities, the global reach of AI-generated content and everyday AI use is bound to further skyrocket, increasing the overall usage quotient of conversational artificial intelligence.
Why Meta AI is a Significant Development for Conversational AI Tools
Meta’s entertainment-based chatbot lays the foundation for a more robust framework for future development.
Meta AI combines various functionalities such as conversational AI, AI search, and AI image generation and also acts as a simplistic AI assistant. The chatbot has been primed to offer concise answers and in an amicable tone to simplify the process of AI use. The integration with Meta’s current applications also displays the ability to scale AI tools and the innate flexibility of Meta AI. While it’s still early for long-term reviews and analyses of features like Gen AI Personas, Meta AI’s role in the firm’s larger Metaverse investments seems to be playing out in its initial stages. As the tech giant generates greater interest in both the general populace as well as corporate clients, Meta AI is bound to grow further, catapulting Meta into the stiff race between companies to create responsible AI that is also commercially viable.
FAQs
1. What language model does Meta AI run on?
Though the exact details such as parameters and context length are not known, Meta AI runs on a customized version of LlaMa 2—Meta’s in-house language model, which it built in partnership with Microsoft.
2. Can Meta AI generate images?
Yes, Meta AI can generate images through its internal image creation protocol titled “Emu.” Meta has trained the model extensively using both image and text-based pairs.
3. When was Meta AI launched?
Meta AI was launched on September 27, 2023, during the Meta Connect 2023 event, which showcased several new products and capabilities developed by the firm.