Meta unveiled its latest AI chatbot, Llama 3, on Thursday, with CEO Mark Zuckerberg touting its impressive capabilities. The new model is said to deliver 'state-of-the-art' performance on industry benchmarks and features enhanced reasoning abilities, as per a company blog post. Meta plans to integrate Llama 3 into its AI assistant, Meta AI, making it accessible across various Meta products, including Instagram.
Zuckerberg expressed confidence in the new model, claiming that Meta AI is now the most intelligent AI assistant available for public use. However, he reassured that current AI advancements do not pose existential risks to humanity. In an interview with The Verge, Zuckerberg emphasized that the technology Meta and others are developing does not reach the level of concern for such risks.
Regarding the openness of Meta's AI models, Zuckerberg hinted at potential limitations. If Meta achieves multimodality, enabling the delivery of results in different media formats like text, images, and video, the company may reconsider making all aspects of its model open-source. Zuckerberg cited image generation as a specific area of concern, especially in the context of potential misuse during events like elections.
While Meta declined to provide further comments on the matter, the company's strategic approach to AI development appears to balance innovation with responsible deployment. As the tech giant continues to refine its AI capabilities, the implications of advanced models like Llama 3 on societal dynamics and ethical considerations remain subjects of ongoing evaluation.
