Search Here

During their annual conference, Tech Crunch reported that NVIDIA CEO Jensen Huang discussed the possibility of human-level AI or artificial general intelligence (AGI) in the next five years. He added it will be greatly attributed to AI hallucinations.

"If we specified AGI to be something very specific, a set of tests where a software program can do very well - or maybe 8% better than most people - I believe we will get there within 5 years,"

From Huang's perspective, however, several complex issues surround the quest for human-level AI. The first is to emphasize the need for clear definitions due to the ethical and existential questions it poses. The concept's potential to surpass human intelligence raises uncertainties about its decision-making and ethical alignment with human values.

This is where AI hallucinations come in.

The NVIDIA CEO, Jensen Huang, acknowledged that soon, if not during the era of human-level AI, these systems will be able to produce responses that sound credible but lack factual basis.

Therefore, AI hallucinations will be a phenomenon in which answers may seem reasonable based on AI's training data but do not reflect reality.

To counter this, Huang proposes an AI-integrated system.

First, he emphasized establishing reliable procedures to guarantee that AI-generated responses are well-researched and based on verifiable data. Huang introduced the idea of "retrieval-augmented generation," endorsing a systematic approach similar to fundamental media literacy. This involves AI systems cross-referencing information from trustworthy sources to provide accurate responses, especially in critical areas such as health guidance.

"The AI shouldn't just answer; it should do research first to determine which of the answers are the best."

Besides human-level AI, the NVIDIA CEO also discussed their AI chipset. As first reported by VCPost, Huang clarifies that no other competitors will still be able to top their offerings.

Tags

Nvidia, Jensen Huang