AI - DOOM AND HOPE - A TALE OF TWO WORLDS

N°09

This edition is inspired by 2001: A Space 0dyssey*

"HAL 9000: Let me put it this way, Mr. Amor. The 9000 series is the most reliable computer ever made. No 9000 computer has ever made a mistake or distorted information. We are all, by any practical definition of the words, foolproof and incapable of error."

(Click on the photo to access the video.)

Dear readers of FWIW, my journey is this time more algorithmic than physical. AI is making my travel binary while I?wait for quantum to beam me up the Star Trek way.?

To start with, let’s remember that since the dawn of times the real driver of innovation in technology has been the desire to influence the outcome of wars and to advance in the art of conquering. In the nearest past, the origin of the Internet, e-mail, drones... was each time a resolve to improve the art of WAR. Now, AI is no different.??

The roots of Artificial Intelligence can be traced back to philosophers who attempted to describe the process of human thinking as the mechanical manipulation of symbols. The invention of the programmable digital computer in the 1940s was a major milestone in the development of AI. The modern history of AI begins with the development of stored-program electronic computers. The maturation of AI happened between 1943 and 1952, with the development of artificial neurons, updating rules for modifying the connection strength between neurons, and machine learning.?

In 1950, Turing proposed the Turing Test for determining a machine's ability to exhibit human-like intelligence. This seminal idea paved the way for?Artificial Intelligence.?

(Click on the photo to access the video.)

In very general terms (check your ChatGPT for deeper insight!) you have?large language models (LLM), i.e., artificial neuronal systems that learn and predict text structure and sequence. LLMs will develop at a much faster pace once Quantum Learning becomes mainstream.?

Regulatory:

OECD or UNESCO are starting to design international regulatory frameworks.?Right now, the AI business is regulated in our lovely western civilised world as follows:?

U.S. Approach:?

• Decentralised: The U.S. approach to AI risk management is?highly distributed?across federal agencies. These agencies often adapt to AI without new legal authorities.?
• Non-Regulatory Infrastructure: The U.S. has invested in non-regulatory infrastructure, including:?
• A?new AI risk management framework;?
• Evaluations of facial recognition software; and?
• Extensive funding for?AI research.?

EU Approach:?

• Comprehensive Legislation: The EU’s approach features a more?comprehensive range of legislation?tailored to specific digital environments.?
• EU AI Act: The EU plans to place new requirements on?high-risk AI?in socioeconomic processes, government use, and regulated consumer products with AI systems.?
• Transparency and Influence: Other EU legislation enables more?public transparency and influence?over the design of AI systems in social media and e-commerce.?

Alignment, Misalignment:

• Conceptual Alignment: The U.S. shares a conceptual alignment with the EU on risk-based approach, trustworthy AI principles, and the importance of international standards.?
• EU-U.S. Collaboration: The?EU-U.S. Trade and Technology Council?aims to align efforts.
• However, regarding specific AI applications—especially those related to socioeconomic processes and online platforms—the U.S. and EU are on a path to significant misalignment.?

Some people claim that the Chinese AI regulatory framework is actually rather flexible and comprehensive... I wouldn’t know and do not think it would ever be easy for foreigners to find out.?

"I know I've made some very poor decisions recently, but I can give you my complete assurance that my work will be back to normal." - HAL

But let’s look at the politics and my takeaways

In Spain

We want?to create LLMs à l’espagnole. Considering that the government is not yet out of the doldrums after the Catalan amnesty, why not?create an AI with tortilla de patatas flavour, knowing that one of the key elements for the success of any of these adventures is an open model. The more the merrier in terms of data.... well, isolation is not really the way forward. ?

In France

Jupiterian France wants to organise an AI summit at Versailles, and hence has launched Mistral with public and private money to compete with the billions?of Microsoft and the Silicon Valley kids. ?

Bear in mind though that 3 factors make it so difficult to catch up with Open AI, namely (i) the amount of money that has been thrown into developing its products, which is in the billions; (ii) that key data collection companies have been acquired and feed into the existing LLMs; and (iii) that there is a clear competitive first-mover advantage... Well, Mistral may well carry its wind of the season name as a harbinger of success but it’s too early to tell.?

In fact…

...regulators worldwide should get their act together jointly for the result to be sensible and useful. Let’s remember Kubrick’s choice for his 2001 soundtrack: Ligeti’s Lux Aeterna, an example of micropolyphony that has sixteen individual parts converge into an... unintelligible tone cluster.?

Astronaut David Bowman to HAL 9000: “Unless you obey my instructions, I shall be forced to disconnect you.”?

But the double face question: Is AI going to kill the human element and indeed millions of jobs in the process?

Well, no, in my humble opinion!?

Obviously, there’s big hope for Life Sciences, and fear among content creators and publishers. But, overall, this is again a war of money and engineering. Indeed, while a Michelin-star chef is unlikely to lose his job to AI, certain paralegals had better look to Hollywood for alternative career options... but, hey, this Next Big Thing is not going to do away with the job market as we know it.?

The answer is going to be in the implementation of the regulatory frameworks described above. These have the potential to create competitive advantages, but when I look at the EU’s recent Digital Services Act (DSA) and Digital Markets Act (DMA) and their challenges, my dear reader, I would say?that the proof is in the pudding.?

Remember that when Astronaut David Bowman aka Dave takes out HAL 9000, the starship’s mission fails, such as the latter predicts were he to be disconnected.?

And, finally, it is interesting to conclude that given the growing number?of open war conflicts in the world we can be positive about the fact that innovation and leadership on Quantum will be with consumers soon.?

*“2001: A Space Odyssey” warns us to approach AI with caution, emphasizing the need for responsible development, transparency, and thoughtful consideration of its implications.

Please watch the interview with the writer of Arthur C. Clarke talks A Space Odyssey and artificial intelligence, 1968?