# The True Nature of AI | Hazem Ali | TEDxEJUST

Source: https://www.youtube.com/watch?v=3_Gr5Lx81Ns
Recap page: https://rapidrecap.app/video/3_Gr5Lx81Ns
Generated: 2025-12-12T18:11:19.74+00:00

---
## Quick Overview

The speaker, Hazem Ali, explains that the true nature of AI, particularly Large Language Models (LLMs) like GPT, is not about replacing human intelligence but about augmenting it through engineering principles, emphasizing that current AI systems, despite their capabilities, still lack genuine consciousness or reasoning, and this distinction must be understood by engineers and users alike.

**Key Points:**
- The speaker, Hazem Ali, distinguishes between current AI capabilities and true human intelligence, noting that modern LLMs like GPT still lack consciousness or independent reasoning (0:32, 2:54).
- He contrasts academic research in AI, which focuses on capabilities like SQL queries or infrastructure execution, with practical, real-world application (3:41, 4:46).
- Ali argues that for AI to be truly useful, it must be 'connected' to the real world, unlike current systems that only execute instructions without understanding context (4:51, 5:25).
- He highlights that while AI is rapidly advancing, particularly with models like GPT-4, it is still limited by its training data and cannot truly 'think' or make judgments beyond its programming (2:55, 4:14).
- The speaker suggests that engineers must understand the difference between AI's ability to execute tasks (like generating code or data) and true comprehension, referencing concepts like prompt injection as a limitation (4:46, 10:34).
- Ali concludes by stating that the goal should be to use AI as an assistant to enhance human roles, such as Project Manager, rather than expecting it to replace the human element entirely (8:46, 13:34).
- The session includes a discussion about upcoming topics, such as advanced concepts in AI and a potential workshop on prompt engineering and security (5:33, 8:52).

![Screenshot at 0:04: Speaker Eng Hazem Ali introduces himself during the TEDx EJUST event titled 'Beyond The Boundaries 2025'.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-00-04.png)

**Context:** Hazem Ali presents his insights on the current state and limitations of Artificial Intelligence, specifically Large Language Models (LLMs), during a TEDx event themed 'Beyond the Boundaries' in 2025. He addresses the common misconceptions people hold about AI, particularly concerning its ability to reason, acquire consciousness, or replace complex human roles like engineering or project management.

## Detailed Analysis

Hazem Ali discusses the nature of Artificial Intelligence, particularly Large Language Models (LLMs) such as GPT, contrasting their current capabilities with genuine human intelligence. He asserts that while these models are powerful tools for tasks like executing SQL queries or infrastructure commands, they fundamentally lack consciousness or true reasoning, which is a key limitation that people often overlook (0:32, 2:54). Ali stresses that the current trend involves engineers and users treating AI as capable of complex thought when it merely processes information based on training data and explicit instructions. He points out that the real value of AI lies in its augmentation of human ability, not replacement, citing the difference between academic research (e.g., building models) and real-world impact (4:46, 8:46). He uses the example of prompt injection to show that AI executes instructions literally without contextual awareness. The speaker mentions that practical application requires the AI to be 'connected' to the real world, not just perform rote tasks. He also touched upon future topics for deeper dives, including advanced AI concepts, security for agents, and prompt engineering principles, which he suggests will be necessary to integrate AI effectively into roles like engineering and project management (5:33, 13:34).

### Introduction and AI Limitations

- Introduction of Hazem Ali
- Discussion focuses on the difference between current AI (LLMs like GPT) and true human consciousness/reasoning
- Current AI executes instructions but doesn't 'think' (0:32, 2:54)

### Academic vs. Practical AI

- Contrast between academic research (building models, running queries) and real-world application/impact
- Real-world application requires 'connection' to reality, not just instruction following (3:41, 4:51)

### Future Implications and Roles

- Mention of future feasibility studies (2025-2040)
- Discussion on how AI should augment roles like Project Manager and Engineer, not replace them (8:46, 13:34)

### Key Concepts in AI

- Mention of Prompt Engineering and Security Agents
- Distinction between academic knowledge and practical application in the field (10:34, 13:58)

![Screenshot at 0:01: Title slide for the TEDx EJUST talk, 'Beyond The Boundaries', featuring Eng Hazem Ali and the year 2025.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-00-01.png)
![Screenshot at 0:14: Speaker referencing trending AI topics like Chat GPT and LLMs.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-00-14.png)
![Screenshot at 0:43: Slide indicating the topic is divided into two main sections.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-00-43.png)
![Screenshot at 2:14: Speaker defining the concept of 'limitations' in AI systems.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-02-14.png)
![Screenshot at 4:00: Speaker discussing the concept of a judge making a decision based on law, not emotion.](https://ss.rapidrecap.app/screens/3_Gr5Lx81Ns/00-04-00.png)