Meta's recent release, the large language model Llama 3, has fueled the development of "Meta AI," a new chatbot embedded across various Meta platforms. But how does it measure up against other conversational AIs? The verdict is mixed.
Accessible for free on Meta.ai, Instagram, Facebook, WhatsApp, and several other platforms, Meta AI seems to prioritize convenience over performance. Its responses are often curated from web search results, with a tendency to provide generalized information rather than excelling in any particular area.
Meta's CEO, Mark Zuckerberg, boldly anticipates Meta AI to become "the most used and best AI assistant in the world." However, early reviews reveal a chatbot that, while functional, may not quite live up to such grand expectations.
In a series of informal evaluations, Meta AI's performance was compared to other conversational AI models. The results showed that while Meta AI offers factual and up-to-date information, its responses tend to lack depth and originality.
Queries on current events and social media trends yielded search-like responses, often leading to external sources without providing substantial insights. For instance, when asked about TikTok trends, the response was generic, failing to offer recent trends.
Similarly, inquiries for historical and contextual information sometimes resulted in irrelevant or outdated responses. Meta AI's performance in offering medical advice and addressing sensitive topics was acceptable, albeit lacking in sourcing and depth.
Overall, while Meta AI serves as a convenient tool for casual inquiries, its reliance on web search results and its limited depth may leave users seeking more profound insights dissatisfied. As Meta continues to refine its AI capabilities, one can hope for improvements that align more closely with user expectations.