AI and Blockchain: A technological revolution

Towards a collaborative future: Optimistic perspectives on AI and Blockchain

Welcome to our series exploring the potential futures of Artificial Intelligence and Blockchain. Each article delves into a future envisioned with optimism, reflecting the endless possibilities offered by these technologies. We are navigating in the realm of foresight, where imagination meets innovation.

DALL.E-3, Gaspard Tertrais

Generative AI, is already an integral part of our daily lives. From virtual assistants like ChatGPT to open-source systems like Meta’s LLaMA, these technologies shape our interaction with the digital world. However, one question remains: how can we ensure these advancements benefit everyone and reflect our societal values?

Section 1: Generative Artificial Intelligence, explained

DALL.E-3, Gaspard Tertrais

GenAI: What is it?

GenAI (Generative AI) is a type of artificial intelligence that can create new content, such as text, images, or music. It uses deep learning and neural networks to analyze vast amounts of data and identify patterns.

Unlike traditional AI that simply identifies or classifies elements, GenAI generates original creations that mimic the style and content of its training data. Language models like OpenAI’s GPT-3 or Meta’s LLaMA, trained on huge text corpora, can produce coherent and relevant responses in conversations, summarize texts, answer questions, and even write stories.

How does this digital magic work?

These GenAIs are like digital sponges. They absorb an incredible amount of information — books, articles, websites — and use all this data to understand how to speak, write, and even think. It’s akin to learning a language by reading every book in a library.

Algorithmic bias: A significant challenge

In the world of AI, not everything is perfect. One of the big concerns is “algorithmic biases.” Essentially, if AI learns from biased data (like discriminatory texts or stereotyped images), it will reproduce these biases in its creations. It’s like a child unknowingly learning bad manners. Therefore, monitoring what AI learns is crucial to prevent the spread of prejudice or misinformation. But who decides what data is biased or not?

Fine-tuning: educating AI post-training

Then there’s “fine-tuning,” akin to AI’s continuous education. After AI learns from a huge amount of data, developers can refine it for specific tasks or to make it more accurate. It’s like giving AI private lessons to learn to do certain things better or to correct mistakes it might have learned during its initial training. This process is essential for quality AI, but who decides its education, its political or moral direction?

The impact of generative AI in our lives

This type of AI isn’t limited to chatting or writing texts. Its implications are vast, such as:

  • In healthcare: It can help discover new drugs by analyzing millions of scientific data.
  • For businesses: It optimizes processes, responds to customers, and even aids in decision-making.
  • In artistic creation: It can compose music, create artworks, or write scripts.

Closed AI vs. open-source AI: A matter of freedom

In the AI universe, there are “closed” versions like OpenAI’s ChatGPT (not so open since Microsoft’s billion-dollar investment in this pseudo non-profit initiative) and “open-source” versions, free and accessible to everyone, like Meta’s Llama. The difference? It’s like having the recipe for a delicious cake: either it’s kept secret, or it’s shared so everyone can cook and improve it.

Imagine the world of tomorrow, in 10, 20 years, where we all use generative AIs for all sorts of tasks. Would you prefer to know that the data of these AIs are open to everyone, or that these AIs are closed and controlled by a small group of people who would decide how these systems (and therefore you!) should think?

The current greatest risk is not that these AIs become “conscious” and take over, but knowing who controls these AIs. I highly recommend the excellent debate between Yann LeCun (CTO of Meta) and Brad Smith (VP of Microsoft) on the topic of an open and democratic AI:

https://youtu.be/KLevsx_LaBo

AGI: The future of AI

AGI, or Artificial General Intelligence, is the next level. Imagine an AI not only capable of doing many different things but also of learning and understanding them like a human. It’s the ultimate dream of AI technology: a system that could think, learn, and act autonomously in almost any field.

Towards singularity and alignment

  • The technological singularity: It’s the idea of a future where AI becomes so advanced that it can improve itself and thus completely transform our world. It’s both exciting and terrifying because such misaligned AI could cause very serious problems, with some even thinking of human enslavement or extinction, like the Skynet of Terminator!
  • AI alignment: It’s the challenge of ensuring that these super-intelligent AIs understand and respect our human values and goals. It’s like raising a prodigy child who must learn to use their talents wisely.

By better understanding generative AI, we begin to see its enormous potential and the way it could transform our future. It’s a fascinating field, full of possibilities, that deserves our close attention.

Section 2: Blockchain explained

DALL.E-3, Gaspard Tertrais

Blockchain: What exactly is it?

Imagine a huge notebook that everyone can read, but in which no one can erase or modify a page once it’s written. Each new page (or “block”) contains information about exchanges or transactions, and once it’s complete, it’s sealed and attached to the previous page. Thus, they form a long chain of pages, hence the name “blockchain.”

The big ideas behind Blockchain

  • No Boss: Instead of being stored on a big central computer, this chain of pages is spread across thousands of computers worldwide. No one “owns” the blockchain, much like a guest book in a public library.
  • Open to All: Everyone can take a peek into this notebook and verify what’s written. Everything is transparent and open. Super Secure: Thanks to complicated encryption techniques (like sophisticated digital locks), it’s extremely difficult to cheat or alter what’s been written.

When Blockchain meets real life

The magic of blockchain isn’t limited to cryptocurrency exchanges. It can transform many things in our daily lives:

  • Voting with confidence: Imagine elections where every vote is indelibly and transparently recorded.
  • Simplified home buying: No more endless paperwork to prove who owns what.
  • Safer medications: Precise tracking of medications, from their creation to their arrival at the pharmacy.
  • Authentic products: No more counterfeits, thanks to transparent tracking from production to delivery.

What if we mixed AI and Blockchain?

It gets really exciting when we combine AI and blockchain. Imagine an AI that can read and understand this immense chain of pages to help us make intelligent decisions, all while being monitored by the transparency and security of the blockchain. This could mean major advancements in fields like finance, health, and many others.

Section 3: When AI meets Blockchain — An alliance of giants

Now, let’s get to the heart of the matter: why AI and blockchain are like two pieces of a giant puzzle of innovation. It’s like two superheroes deciding to combine their forces!

DALL.E-3, Gaspard Tertrais

Why this super alliance?

  • A powerful duo: Imagine AI as a super-fast brain and blockchain as an infallible memory. Together, they form a dream team: AI analyzes and creates, while blockchain ensures everything is secure and transparent.
  • Stronger together: AI can be super intelligent, but it needs reliable data. Blockchain is like a guardian that ensures AI’s data are neither tampered with nor manipulated.
  • Technological democracy: With blockchain, we avoid AI being the toy of a few large companies. It’s like giving control of this super-powerful intelligence to everyone, not just an elite.

The potential of this magical union

  • Total transparency: Thanks to blockchain, we can track every decision of the AI like footprints in the sand. It’s reassuring because we know where the decisions come from and how they are made.
  • The economy of the future: Imagine a world where everyone can contribute to AI and be rewarded for it, much like in a video game where each player earns points for their participation.
  • Innovations everywhere: This combination could transform everything, from finance to medicine, to education. It’s as if each sector received a technological superpower.

A huge impact on society and economy

  • Technology closer to us: By integrating blockchain, AI becomes like a trusted friend who always acts in our interest, transparently.
  • Power to the people: This alliance gives us control over our own data and how AI is used. It’s like taking back the reins of technology.
  • Towards a fairer future: By making access to AI more equitable, we ensure that the future is not just bright for a few, but for everyone.

Selecting data with blockchain: a new era of transparency

Blockchain can revolutionize how we choose data to train our AIs. Using its transparent and immutable nature, we can create a system where each dataset is recorded, verified, and validated. This means AI developers, researchers, and even the general public could track the origin, nature, and quality of the data used for training.

Infallible traceability for AI fine-tuning

Fine-tuning is a key step in AI development, where the model is adjusted for optimal performance. Here, blockchain can play a crucial role by recording every modification, choice, and data source used in this process. This offers total transparency, allowing us to understand how and why the AI evolved in a particular way.

Multiple benefits of this approach

  • Increased trust: By documenting every step of the AI development process, blockchain strengthens public and professional users’ trust in generative AIs.
  • Bias reduction: With clear traceability, it is easier to identify and correct potential biases in datasets.
  • Improved governance: This method allows for democratic governance of AI, where decisions about data and adjustments can be made transparently and informed.

Conclusion: Towards a future co-created with AI and Blockchain

Generative AI has unveiled a world where machines create, opening up new possibilities. Blockchain acts as an open ledger ensuring security, transparency, and fairness — it underpins cryptocurrencies and enables reliable, democratic technology. The convergence of AI and blockchain marks the start of a new technological era in which AI becomes our partner, with blockchain providing incorruptible oversight.

[Article written on October 26, 2023, by Gaspard Tertrais with the support of the Open AI GPT-4 algorithm for approximately 20%. Images created with DALL.E-3, 2023]

--

--