Multimodal LLMs advance AI capabilities: Exafluence’s Rout sees path to AGI

Exafluence’s Director of Data Science, Malaya Rout, emphasizes the significant evolution in AI through multimodal large language models (LLMs), which combine text, vision, and audio to create more human-like intelligence.

This shift from traditional text-only models to multimodal LLMs has expanded AI’s capabilities beyond language processing to include sight and sound, enabling applications such as content moderation that flags plagiarism, explicit content, toxic behavior, and legal compliance issues.

“The shift has made AI more human-like by embodying a rich interplay of sight, sound, and language,” Rout explains, highlighting the enhanced sensory integration.

He further questions, “Can I say that here we have an AI that thinks, sees, hears, speaks, and reads? Haven’t we moved multiple steps closer to AGI (Artificial General Intelligence) with this?” This suggests that multimodal LLMs are a critical step toward achieving AGI, where AI systems possess generalized cognitive abilities similar to humans.

Rout’s background, including his leadership role at Exafluence and experience with major firms like TCS and Verizon, lends credibility to his insights on the future trajectory of AI.

As multimodal LLMs continue to evolve, their integration of multiple sensory inputs is expected to drive AI closer to true general intelligence, transforming industries and redefining human-computer interaction.

Exafluence’s Director of Data Science, Malaya Rout, highlights the transformative shift from text-only to multimodal large language models (LLMs), which integrate sight, sound, and language, advancing AI capabilities and bringing the technology closer to Artificial General Intelligence (AGI).