How dangerous is AGI ?

Envision a future where AI robots possess human-like intelligence. Sounds like what happens in one of those sci-fi movies, doesn’t it? But hear me out, you won’t believe what Sam Altman, the CEO of OpenAI has to say about AGI because he believes human-level AI is coming.

But what exactly is AGI? How is it different from the AI we use today? How dangerous is AGI? In this blog post, let’s look at AGI(Artificial General Intelligence) and its implications to the world.

AGI will change the world much less than we all think, and it will change jobs much less than we all think”

 — Sam Altman

What is AGI?

AGI, or Artificial General Intelligence, represents a super-advanced AI. Unlike specialized AI (Artificial Intelligence), such as ElevenLabs for audio generation and Jasper for content writing, which are designed for specific tasks, AGI mirrors human intelligence. Like us, it can understand, learn, think, reason, and solve problems just like humans. However, it goes a step further, because it can also learn and tackle challenges across various fields, possessing knowledge beyond what a normal human brain can contain. Here’s a video that explains AGI in just 5 minutes.

What are the key challenges of AGI?

You may be wondering, yeah right, I guess AGI sounds cool, but how do we even get there? We need both hardware and data. Don’t get me wrong, but our hardware is often limited by computational constraints. Even with NVIDIA’s most powerful AI chip today, it could still take weeks to train large AI models. To bridge this gap and go beyond the wildest dream of traditional microchip-based computation, that brings us to yet another futuristic technology — quantum computing.

Quantum computers have the potential to compute millions of times faster than traditional computing. However, we’re still far from achieving the supercomputing levels we envisioned. 

That’s why NVIDIA is working on hybrid quantum-classical computing. They’ve even developed an open-source platform called CUDA-Q for integrating and programming quantum processing units (QPUs), together with Graphics processing unit (GPUs) and central processing units (CPUs) in one system.

When it comes to data, things are trickier than they may seem. We need not only more high-quality data but also a diverse range of data that can be considered quality knowledge. However, the ethical handling of data and maintaining the right moral principles are crucial because AI can absorb too much data and misinterpret them, leading to data poisoning. Or even worse, AI may generate or perceive data that is not present in reality, which is a phenomenon known as hallucination.

This is often the point where things can take a turn for the worse, and it may just be the day we find ourselves living in an apocalyptic world, just like in the sci-fi movies.

Where could AGI develop the first?

We may not want to imagine it, but when AGI combines with AI robots, things become even more intriguing. Just as the ancient one predicted: when AI becomes so advanced, the next natural step is to give the machine some sort of physical form, by the way, as we explained in this video.

With today’s technology, we’re beginning to foresee the future through projects like OpenAI’s Figure 01 Robots. If you’re familiar with ChatGPT, I meant, if you’re willing to pay a little extra to trade a ChatGPT plus subscription or gain access to AI studio in Microsoft Azure, you’ll have access to most AI models from OpenAI, including LLMs like GPT- 4, or diffusion models like Dall-E, and even Whisper speech recognition (ASR) systems. With the assistance of neural networks, Figure 01 can engage in speech-to-speech reasoning, taking actions concisely. 

Figure 01 is not the only example. China has surged ahead in the robotics industry with the introduction of the groundbreaking humanoid robot, Astribot S1. This advanced, semi-autonomous robot is powered by cutting-edge LLMs. If you want to learn more about AI startup here’s an state of AI startup in 2023.

How could AGI become dangerous?

AGI holds both promise and risk. While it holds the potential to revolutionize technology and innovation, concerns about its control and potential dangers loom large. 

One significant worry is the prospect of AGI becoming too powerful, possibly displacing human jobs and causing economic upheaval. Additionally, there are apprehensions about its potential misuse for weapons or privacy infringements, raising ethical dilemmas and societal concerns.

Despite its transformative potential, AGI also poses significant risks. For instance, imagine an AGI system that manages a household’s energy usage, adjusting heating, cooling, and lighting for efficiency. While this could enhance comfort and reduce utility costs, it raises questions about privacy and dependence on technology. 

Similarly, in workplaces, AGI could streamline tasks, but its over-reliance could lead to job displacement and human skill degradation, posing economic and social challenges.

Furthermore, AGI systems could become targets for cyberattacks, with potentially catastrophic consequences. A compromised AGI system might make autonomous decisions that are harmful to businesses or public safety, given its extensive control over critical infrastructure. This vulnerability underscores the need for robust cybersecurity measures to safeguard against potential threats.

If you think of it, AGI holds the promise of transformative advancements. However, its unchecked development and deployment could lead to unintended consequences. Therefore, carefully considering its risks and proactive measures to mitigate them are imperative to harness its benefits responsibly.

Where could AGI lead us?

So, where could AGI lead us? And how far are we from AGI?

You see, recent technological advancements, such as transformer architecture in deep learning and natural language processing ( NLP ), have propelled us closer to achieving AGI. 

Technological uncertainty 

Imagine a future where AGI-powered self-driving cars eliminate accidents and traffic congestion, improving air quality and safer roads. However, the path to AGI is fraught with technical challenges, ethical dilemmas, and regulatory complexities.

The journey towards AGI remains uncertain due to these obstacles. While we’re making significant strides in AI development, there are concerns about ensuring its responsible and ethical use. Leaders worldwide must come together to establish clear boundaries and regulations to guide AI development and safeguard privacy before potential risks spiral out of control.

Ethical concerns

As we envision a future where AI systems become increasingly capable of solving complex problems, questions arise about our ability to control them effectively. The development of AGI raises important considerations about governance, accountability, and transparency.

Interestingly, the impact of AGI development extends beyond technological advancements. In Europe, the introduction of the AI Act, while not directly targeting AGI, could indirectly influence its progress. 

Regulations 

The regulations and principles outlined in the AI Act may shape the landscape of AI research and funding opportunities in Europe, potentially impacting the pace of AGI development in the region. While the AI Act focuses on regulating specific aspects of AI systems, its broader implications underscore the need for careful consideration of AGI’s societal and ethical implications.

Natural Language Processing with Transformers, Revised Edition

Since their introduction in 2017, transformers have quickly become the dominant architecture for achieving state-of-the-art results on various natural language processing tasks. If you’re a data scientist or coder, this practical book -now revised in full color- shows you how to train and scale these large models using Hugging Face Transformers, a Python-based deep learning library.

Transformers have been used to write realistic news stories, improve Google Search queries, and even create chatbots that tell corny jokes. In this guide, authors Lewis Tunstall, Leandro von Werra, and Thomas Wolf, among the creators of Hugging Face Transformers, use a hands-on approach to teach you how transformers work and how to integrate them into your applications. You’ll quickly learn a variety of tasks they can help you solve.

Check it out on Amazon

Put them together

Because of those technical challenges, ethical concerns, and regulations, this made the road to AGI remains uncertain. As this new age of AI unfolds with rapid advancements, the extent of its impact is still unknown. It is crucial for global leaders to convene and establish boundaries for AI development and privacy protection before potential risks become unmanageable.

But here’s my question : imagine one day when AI becomes so good at resolving problems for humans, are we confident enough to ensure its control?

Looking forward

I plan to try different methods and document my journey here and on YouTube. So, please feel free to follow me here on Medium and subscribe to my YouTube channel if you want to follow my journey. Let me know your thoughts in the comment section. Stay tuned, and see you in the next one!