Two years ago in these pages we explored how Artificial Intelligence (AI) had burst into prominence with Chat GPT. It seemed amazing that it could carry on a conversation or write essays on an endless variety of topics. Back then the latest versions were just beginning to perform more complex tasks like writing computer code. In that article we outlined the history of AI and looked into what the future might bring. We noted that it has the potential to bring about positive changes in our world, but there are also potential dangers inherent in this technology. On the positive side, AI can automate many tasks to increase efficiency and productivity. The downside of that is the millions of people who would lose their jobs in just about every area of society. We also observed that relying too much on AI systems could make humans overly dependent on technology, and they would be helpless when a system fails or is compromised.

Well, it has been only two years, but it looks like that future is almost upon us. AI systems have already started surpassing humans in a number of areas including coding, math and medical diagnosis - and they are getting better every day. The original models struggled with complex reasoning problems, but that has been solved by increasing computing power to deepen their data base. Today’s most advanced systems are starting to approach what is called Artificial General Intelligence (AGI). They are trained to take an additional step before giving a response. In other words, they pause for “reflection” to “think” about what they are going to say. This sounds disturbingly human, so let’s explore the rapid growth of AGI and how it could affect our lives.

AGI refers to a level of artificial intelligence that has the ability to understand, learn, and apply knowledge across a wide range of tasks, at a level equal to or exceeding that of a human being. Unlike most current AI systems which are specialized (playing chess, recognizing faces, analyzing radiology, generating text, etc.), AGI would be able to communicate and interpret meaning on countless topics in human-like ways. It could reason abstractly, solve problems, adapt to new situations and “think” on its own. For example, two years ago AI computer coding was aimed at speeding up human coders, not replacing them. Today AI already does most of the actual coding, and the job of the human engineers is primarily to supervise the AI systems. If AGI is fully developed – meaning an AI system that can understand, learn, and perform any intellectual task a human can – then most jobs involving cognitive labour would be at risk.

Professions that involve predictable logic and text processing are likely to be the first ones that are completely replaced by AGI. These include accountants, bookkeepers, legal researchers, customer service representatives, and middle management of businesses. Other professions like doctors, lawyers and financial advisors will still be controlled by humans, but much of their work will be done by AGI. This means that many of the people who now assist these professions will no longer be necessary. On the bright side, jobs that require physical presence, empathy, trust, ethics, and complex social interaction are the least likely to be affected by AGI. This includes teachers, therapists, psychologists, healthcare workers, and skilled tradesmen like builders, electricians and plumbers. AGI may aid them, but it will not replace them. On the other hand, truck drivers and taxi drivers are already being threatened by self-driving cars and trucks. It is hoped that journalism, art and entertainment will still rely on the talents of individuals. But much of our music, films, novels and news is already being produced by AI, and AGI threatens to make them even more automated.

Another big factor in the rise of AGI is that it requires massive amounts of electricity to run its gigantic data centres. In a world where we are scolded to conserve energy for “Net Zero”, our largest tech industries are gearing up to vastly increase their use of electricity. One example is Amazon’s plan to build 30 huge data centres on a 500 hectare site packed with millions of specialized computer chips. It will have several hundred thousand kilometres of fibre-optic cable that will connect the entire complex into one giant machine to service artificial intelligence. This site will consume 2.2 gigawatts of electricity - enough to power a million homes. It will also use millions of gallons of water to keep the chips from overheating. Google, Microsoft, Apple and Meta (which owns Facebook, Instagram and WhatsApp), are also building data centres that are as large or larger. Experts predict that as much as half of all new power demand by 2030 will be to run AI data centres. That is only five years from now. Where will this extra power come from, and will that be good for the environment? 

Where is this all heading? Who knows? Will AI be a wonderful tool that makes our lives more interesting and productive, or will it rob our societies of their humanity? Two years ago, more than 1,000 technology leaders including Elon Musk, published an open letter which warned that: “AGI developers are locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one – not even their creators – can understand, predict or reliably control. This presents a profound risk to society and humanity.” Just this month Musk made an updated assessment of where AGI is, and where it is going. “AI is already much smarter than most humans, but not yet smarter than the very smartest humans on a given topic. But it will probably be smarter than any single human at anything in less than two years, and smarter than all humans combined in about five years. We must make sure that AI is maximally truth-seeking.”

The ultimate dystopian nightmare sees AGI becoming so autonomous that it develops the “human” capabilities of scheming and deception. It could then re-write its own code and take over the world. I would like to think that this is not possible, yet our governments and the big tech companies are going full speed ahead with few guard rails. By continuing to move toward reliance on AGI systems that handle most of our cognitive tasks, we could risk losing much of what makes our lives human. This does not seem like a good idea to me. What do you think?