Human future in an AI-ascendant world

As AI continues to evolve, the pressing question is: will its capabilities evolve and encroach upon us to the point that AI will “take over” humans? In this post, I contemplate the trajectory of AI’s dominance, and present my viewpoint on this question.

“AI”

Will AI take over humans? Let’s start off with some definitions. The key term here is AI or artificial intelligence.

Encyclopedia Britannica defines AI as:

the ability of a digital computer…to perform tasks commonly associated with intelligent beings. The term is frequently applied to the project of developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason…generalize, or learn from past experience.

The phrases of note are: intellectual prowess and ability to learn or reason. This is what distinguishes AI, from prior information technologies.

Artificial Intelligence (AI) has existed for decades, but recent developments have made it more concrete and relevant through the rise of LLMs such as ChatGPT and Gemini. This exposure has popularized the term ‘AI’, imbuing it with a sense of infinite possibility.

Hardware and software have advanced a lot in their own right. When we group those advances with AI, we get a new “peoples” definition for AI. It alludes to the zenith of what is achievable in computing evolution, representing intelligent computing in its most holistic sense. This is the definition I will be using.

The other key term is “take over”. AI is already being used in lots of areas like medicine, predictions, image recognition etc. Here, AI is being used to help humans. A “takeover” is something deeper, and more idiosyncratic. I’ll take the concept of “lives and livelihoods” and break it down to components we feel personally: social interactions and work.

Digitization has already revolutionized social interactions — for better or worse. From real friends to pen-pals to “Facebook” friends, technology has changed the reach, format and quality of social communication.

Will AI add to it? AI is already being used for various types of auto-complete, reminders, suggestions which can finesse our interactions. At a bigger level, tools like ChatGPT can remember our conversations and use that as added context for future queries. This is far more powerful influence. LLMs get better every day and with and with added context, they can address broad, life-specific questions e.g., which career should I pursue or even more private questions. In my experience, the contextual answers are quite good — and will only get better.

We could have an LLM become our best friend, counsellor, guide, assistant, sensei or even more.

Will this make the loneliness epidemic more acute? Will it increase dependency on tech, in a world already reeling with effects of social media addiction? Probably, but not in a “takeover” sense. Many of these issues are pre-existing and not the creation of AI. If anything, in the era of information overload, AI could help us streamline our communications and act as an intelligent, non-human — yet “human”, sounding board for our thoughts, ideas, or dilemmas.

Work

Jobs

Many job cuts are ascribed to AI. How much we trust such justifications from organizations is up to us. While there may be elements of truth, I don’t feel AI applications have been so capable, refined and tailored to an organization to have caused the job cuts and shutdowns of the last 2 years, even if they were done anticipating the long-term AI-based future.

Job losses in the Western world are due to a multitude of factors. After COVID lockdowns, economies and supply chains have struggled to recover. Cost of living crisis, weak demand, quantitative easing, recession, stagflation — it’s a vicious economic word-soup out there.

Social trends complicate things further. Self-righteousness emanating from woke ideologies is one. The second, more powerful one, is the impacts of arrangements like Work From Home. Many people claim WFH makes them more productive, but its impact on overall organizational productivity, and the wider economy is debatable. Then there is “quiet quitting”, social activism by employees etc.

As the economic crunch hits, pressure on producing real earnings and real profits increases. Governments are being forced to downsize. There is already top-down pressure on jobs. The new sensitives and reconditioning of people is now creating bottom-up pressures also. Remote working at the same time is getting more acceptable. 2 + 2 = 4. Employers are recognizing that if a job can be performed from home, it might also be outsourced to another country where labor costs are lower, leading to offshoring. AI of course can take this one step further and eliminate the need for human labor in numerous job functions.

While the growing capabilities of AI provide a tempting alternative to employers in today’s tight economic environment, AI is not decimating jobs at the moment: socio-economic conditions are.

At this point in time, AI is not taking over humans, but current conditions are clearing the field and inviting AI to take over. In the future — for sure, AI will replace workers, reduce jobs, drastically alter and eliminate job functions and completely upend the nature of work.

Now, technology has a long history of displacing jobs. Paradoxically, it’s also created jobs for people developing these technologies. With AI, we are seeing demand for AI/ML skills increase however, any rise will be offset by rise of low code and no code tools, many powered by AI. AI is an intelligent technology. While dumb(er) technologies required human development, and created more jobs for humans in the knowledge economy, chances are that future AI advances will also be done by AI.

Knowledge Economy

In the Industrial Revolution, we had mechanical machines, physical products and physical skills. As we transitioned to the Information Revolution, machines became more automated, products became more digitized, and physical skills conceded space to knowledge skills.

With the Internet, information, which used to be the preserve of a select gatekeepers (mostly academia and media), became democratized. It was more accessible than ever before. Information systems could also yield amazing amounts of data. The technology revolution of the 80s onwards spurred significant demand for knowledge-work.

The essence of knowledge-work is gathering information, digesting it, and using it to make decision and/or take action.

The hard part in solving a problem is gathering relevant information, figuring out what’s important, and then acting. AI has stand-out capabilities in the first two aspects — and is getting better.

AI is set to dominate the information gathering space, already providing instant responses to a wide array of subjects. Search, the poster-child of the Information Age, gave us places we could seek to arrive at an answer. AI provides the answer directly. Driven by convenience, answer quality, or peer influence, our reliance on AI for answers will only increase.

The diagnostic and predictive capabilities of AI are boundless in both scope and potential benefits. Smart watches for example can predict certain health conditions well in advanced, based on ML analysis of billions of data points it’s collected.

The decision-space will still rely on human judgement. This is because AI is not there yet. Many AI suggestions are unsatisfactory or incorrect even. However, as our acceptance of AI, and dependence on it increases, and as AI gets better, we may eventually make concessions in this space too.

We can expect decisions of low(er) importance, and where actions can be completed digitally, to be delegated to AI — e.g., auto-generating a tailored report to a recipient. For more involved matters, AI will start to play a bigger role, driven (again) by convenience of the power of defaults. Risk is a factor decision-makers want to control. Humans can sense subtle risks from the environment. While AI cannot replicate that, it offers an equivalent facility — a model trained on 1000s of scenarios than can generate a comprehensive list of commonly observed risks. The quality of human intuition is (semi) offset by the quantity of machine output.

One area where AI cannot venture is physical tasks. It can’t build a house or provide physiotherapy — until such time as autonomous robots are commercially available and socially accepted. AI’s role is expanding in areas where actions can be performed digitally. However, for tasks that require physical work, or matters of the heart, AI remains irrelevant.

Knowledge Work

Using war analogy, the world we’ve established on information and digital processes is under siege by AI. For some fields, like digital artistry, AI already has air supremacy. For others, it’s moving towards a naval blockade. Content creation is another where the influence of AI is palpable. Digital replaced paper. AI is replacing digital.

Regarding other fields, the prevailing opinion is that AI is unlikely to replace jobs that require creativity, empathy, or complex decision-making skills. While this theory is comforting, I’d like to challenge it. Creative fields are already under siege. Let’s look at other two.

Human touch. Soft aspects are important but only in the context of a corresponding hard aspect. For instance, when an organization introduces a new system, the soft skill of change management becomes vital to facilitate easy adaptation and training. However, if the individuals are already well-informed, the transition may not be as significant, and the need for soft skills may be considerably less.

When the hard aspects become easier or automated, the need for human touch (empathy) will also reduce. The impact of AI, or lack thereof, will depend on what the human touch uniquely solves for.

Complex problems. How do we define complex? A lot of things are nowhere near as complex as they are made out to be. Complexity is often invented, and can even be a related to subject matter competence. So, complexity is arguably a subjective term. Software has already streamlined much of the grunt work — research, analysis, and insights — necessary for solving complex problems. AI will elevate this much further. Chess for instance is a complex game with lots of permutations and strategies. After the arrival of Deep Blue, a milestone in AI evolution, it’s widely acknowledged that a human chess player will never beat a machine again. Elon Musk went as far as saying that the advent of computers has made chess a “simple game”.

The conventional wisdom that AI won’t be able to do softer, creative or complex tasks is misplaced. AI will assume much of the mental labor, particularly tasks involving data, information, and knowledge. With its ability to remember and adapt contextually, it might even evolve into a wellspring of wisdom.

Open AI’s 5 waypoints for AI: conversations, reasoners, agents, innovators and organization, already signal to that future. While AI is still at the first stage, the vision is there, and the capability is fast arriving.

Will AI take over humans? On the two axes we discussed — social interaction and work, or lives and livelihoods in a broader sense, the outcomes align in direction by vary in magnitude. The social impact is unlikely to be much more extraordinary than previous information technologies. Work wise, AI will create radical upheaval.

AI will upend large parts of our existence. It will take over the world we have constructed over the past 50 years, a world founded on the information economy and intellectual labor.

Physical work, the cornerstone of the Industrial Revolution, which preceded the Information Revolution, remains relatively secure for now.

Ironically, this could lead us back to areas where AI has yet to make inroads — the sphere of physical services and human interaction, of hands and hearts — areas we had largely abandoned in the race for rapid technological progress.

By taking over knowledge work, AI will free us to address the real issues facing the planet — environmental degradation, habitat loss, waste. AI will take over humans, but it might be just what we need to save humanity.