By Eli Fathi and Peter K. MacKinnon

Eli Fathi (C.M.) (photo below, at left) is chair of the board at Ottawa-based MindBridge Analytics. Peter MacKinnon (photo below, at right) is a senior research associate in the Faculty of Engineering at the University of Ottawa, and Member IEEE-USA Artificial Intelligence Policy Committee, Washington. D.C.

Image at right, an AI panning for gold, was OpenAI-generated.

Panning for gold upon seeing sparkles in a stream has led to gold rushes time and time again. But where is the source of the gold?  Most panners fail and sometimes a lucky few find a vein. 

The Klondike gold rush of 1896 to 1899 in Canada’s Yukon Territory saw some 100,000 prospectors pour into the territory in search of fortune and fame but also created undesirable socio-economic consequences. 

Today, AI is figuratively panning for gold and finding nuggets of various sizes in the streambed. However, panning is not indicative of the source of the sparkles other than the assumption that it must be somewhere upstream, so at least a direction was known.

This state of AI is referred to as Generative Artificial Intelligence (GAI) which McKinsey estimates “could add the equivalent of $2.6 trillion to $4.4 trillion annually” to the global economy.  

In the past few years, AI has found various nuggets of knowledge and knowhow in the shape of different sized nuggets, some big like Large Language Models (LLMs) and others much smaller like AI tools for financial fraud detection. Yet where are the gold veins dispersing these nuggets?

As with many new commercial endeavours, the key drivers are often based on greed and the fear of missing out, just like those who headed for the elusive goldfields.

Currently, we are seeing massive investments in AI by both public and private corporations (such as Google and Alibaba among big players and Cohere and DeepSeek as startups) as well as by many governments.

Over the past few weeks, there has been a steady stream of announcements regarding significant investments in AI, such as US$100 billion from Amazon, €109 billion by France and €200 billion from the EU to name a few. These massive funds are leading to transforming the AI gold rush into an AI gold race. 

In emphasizing the point of this competition, the chief technology officer of Denver, Colorado-based Palantir, Shyam Sanka, said last month that “We are in an AI arms race" between the U.S. and China and that it is “Winner takes all.”

These concepts are relevant not only for national security reasons, but also for commercial and economic reasons. They offer great incentive for corporate entities to make large investments in hopes of being the first to the finishing line.

Andrew Ng, a founder of Google Brain, recently spoke of Google opening the door to the application of AI for military use by saying he was “very glad that Google has changed its stance” in deciding to drop its pledge not to build AI systems for weapons.

Given the potential dual use of AI for commercial and national security reasons, there is an increasing impetus for countries to enter the AI race. The ramifications for the winner could be commercial and geopolitical advantages and even dominance, especially if a single nation achieves Artificial General Intelligence (AGI), creating human-level performance or above.

Little monetization seen so far with Generative Artificial Intelligence

Since the start of the latest “AI spring” circa 2015, significant technological success and progress has been made, but with limited monetization especially compared with the investments made to create existing AI products, such as LLMs.

Although there are many promising startup companies that are developing a plethora of new applications with significant commercial potential, monetization of GAI has been anemic so far. The US firm OpenAI wants to raise billions of dollars to do all kind of projects to create more and bigger nuggets to monetize. 

On the surface, the successes with GAI and future success with AGI will give governments and corporations the ability to offer better services to their citizens/customers, improve efficiency and quality, and provide vastly enhanced capabilities. 

This is the first time in human history that the “pot of gold” at the end of the rainbow is truly intangible. Software is massless, frictionless and instantly transferable anywhere in the world. Previous discoveries such as fire, gold and electricity are physical and tangible. 

Although electricity uses electrons for transmission, it requires significant infrastructure of power generation stations and transmission lines to deliver value. Similarly, the internet, although powered by software, is not controlled by one entity but rather has a large interconnected infrastructure to support its operations. 

History suggests that following the technology curve over time provides a good indicator as to the future trajectories of both GAI and AGI. Take for example the initial decoding of the human genome as an analogy. That initiative started in 1990 and took 13 years to achieve at a cost of nearly $3 billion.

Today, the cost to analyze an individual’s genome is now below $100, a staggering decrease to near zero and applicable to a variety of uses. Similar expectations in cost reduction and increasing functionality are envisaged for AI tools, from frontier LLMs to AI Agents and to even a lower cost range of applications. 

If the AI gold rush hits paydirt, especially with the arrival of human-level AI in the form of AGI within the next few years, all bets are off as to what that will mean to humanity. This begs the question as to whether society is ready for such a profound change in our relationship with technology.

To further compound the situation, this immense power will be in the hands of a few winning entities that will both aid and influence people individually and societies as a whole. Meanwhile, even with the massive amount of investment to date, the economic windfall predicted remains elusive. 

R$