"Eliza Labs CEO claims current AI systems may already meet the definition of Artificial General Intelligence (AGI), as coined in 1997. AGI is designed to m
**Artificial General Intelligence May Already Exist: Eliza Labs Founder**
Eliza Lab's CEO Shaw Walters claims current AI systems could meet the definition of AGI (Artificial General Intelligence). In an interview with Decrypt at ETHDenver, he expressed concerns about the security risks associated with autonomous agents.
**Definition and Origins of AGI**
- Artificial General Intelligence, first coined in 1997, is a theoretical form of AI designed to match or exceed human cognitive abilities across a broad spectrum of tasks.
- It was popularized by researchers, including SingularityNET founder Ben Goertzel.
**Current State of AGI and Future Developments**
According to Walters:
- Current leading models already meet his definition of AGI.
- He does not anticipate a single dominant AI system emerging but rather multiple variants due to the complexity of life.
**Work on AI Agents by Shaw Walters**
Walters started working on AI agents during the GPT-3 era. He found that structured outputs were unreliable and compared the process to "putting training wheels on a baby." Improvements came with the launch of GPT-4 in 2023, enabling more reliable responses and action calling.
**Rising Use of AI Agents**
AI agents have become more prevalent, moving from experimental chatbots to persistent systems embedded across crypto and consumer platforms. Examples include OpenClaw, Moltbook, Agentic Wallets on Base, and Fetch.ai's Visa-compatible agents.
**Security Concerns Regarding AI Agents**
As AI advances towards AGI, it behaves less like a predictable machine and more like a fallible human. This makes foolproof safeguards difficult to engineer. Walters warns of potential risks such as prompt injection and wallet compromises. He emphasizes the need for caution and vigilance as AI development continues.
Source: Read Original Article
Post a Comment