2 Whom is AI a Threat?

If we could direct a film showcasing today’s life with its array of modern gadgets, power tools, war machines, and scientific and medical advancements, and then show it to an audience of the late 1930s, it would undoubtedly be considered a science fiction movie.? Forget about flying cars, teleportation, or time travelling.? From this observation, valuable insights can be gleaned regarding the topic of artificial intelligence.? Never mind Terminators, The Matrix, or the recent surge of apocalyptic narratives surrounding this subject.? The reality, most likely, will lie somewhere in the middle ground but still astound by today’s standards.? However, the imminent change will be profound and directly or indirectly, impact the lives of the majority of human beings.? It may not be immediate, but the speed at which it unfolds will catch most by surprise.

As far as our current knowledge goes, we have yet to produce anything remotely close to general intelligence in artificial systems. Machines (or more accurately, advanced probabilistic models) lack sentience, possess no opinions of their own, and cannot initiate actions independently. Furthermore, there are significant challenges in manufacturing humanoid robots with sufficient autonomy to perform simple tasks, let alone possess the type of general intelligence that would pose a substantial threat to labour-intensive work, whether intensive or not.?

And this already indicates who should not be concerned about recent advancements in AI: the vast majority of labour-intensive tasks that are not particularly repetitive in nature (as those have already been replaced by industrial robots) need not fear. Similarly, tasks that heavily rely on emotional intelligence are also safe, as they demand a level of contextual understanding that machines are unable to achieve. Consequently, entire sectors such as nursing, healthcare, hospitality, construction work, agriculture, entertainment, and many more are unlikely to experience direct repercussions.

On the other end of the spectrum, we have highly qualified professionals who age like fine wines, continuously improving over the years. Think of research, architecture, medicine, and engineering disciplines. Ironically, these are the fields most likely to benefit from AI. Numerous examples in both industry and academia support the notion of a symbiotic relationship between humans and machines, enabling humans to achieve incredible feats previously thought unimaginable. This relationship will manifest as either significantly enhanced productivity or an augmented ability to explore a multitude of options and then narrow them down to the best ones, with the ultimate decision resting in the hands of humans. Expect a rapid acceleration of new discoveries and inventions, surpassing anything seen before. Prepare to be amazed.

Finally, there exists a significant segment of qualified professionals, many of whom possess higher education, whose roles involve moderate levels of cognitive effort. While their tasks may have some repetitive elements, they require attentiveness due to the necessary variations involved. Consider, for instance, the rank-and-file accountant whose job may entail answering questions such as:

• Is the expense supported by a valid invoice?
• Does the invoice meet the criteria of validity?
• Are all items on the invoice eligible as valid expenditures?
• Does the invoiced amount from the supplier align with our monthly expectations?
• Which cost centre should the item be allocated to?
• Are we staying within budget?
• Was the invoice issued in a different currency?
• Which exchange rate should be applied to a foreign receipt?

It may surprise you, but with the provision of contextual information to modern AI systems, the aforementioned questions could be answered by a machine.

Moreover, within the services industry, there exists a broad range of jobs that could potentially be entirely replaced by a suitably configured and integrated AI system. Let us pause for a moment because this sentence encompasses two crucial elements.

Properly Configured AI

What we are actually referring to here is a well-configured or, in technical terms, a trained model. This implies providing sufficient contextual information about a specific human activity to enhance the AI's ability to generate accurate responses (you may already be aware that many publicly available AI systems occasionally produce absurdly incorrect answers).

Returning to the previous example involving the accountant, the AI engine must be equipped with knowledge of the company's policies, relevant cost centres across different departments, and other contextual details specific to the accountant's workplace. Anyone attempting to utilise an off-the-shelf AI engine without employing specifically trained models for their domain or type of problem is likely to dismiss them as ineffective. Beware, as your more astute competitors will make the effort to train the models and achieve what you may have deemed impossible.

This trend is paving the way for a new category of job opportunities. Take a moment to search for the term 'AI trainer jobs' and you will find a wealth of positions currently available. A brief search on two job seeker portals yielded 67 open positions. As time progresses, this number is poised to exponentially increase.

Integrating AI Systems

We previously posited that achieving truly autonomous general intelligence robots, whether humanoid or not, remains beyond our current reach. However, in the absence of such robots, the true utility of AI systems lies in their ability to access the vast amounts of non-public data housed within enterprises. Consequently, efforts must be made to integrate newly trained AI models with corporate systems. It is safe to say that enterprises that have embraced modern API architectures will possess a significant advantage in this regard.

Much is being said about a new generation of intelligent chatbots, and this is indeed the low hanging fruit everyone can understand. However, it is only through deeper integration that bots can truly distinguish themselves by incorporating real-time, relevant data into conversations.

Nevertheless, this is merely the tip of the iceberg. Numerous scenarios exist where humans review documents presented to them and make decisions based on that information. Consider scenarios such as passport renewal applications, tax returns, benefits claims, job applications, and the list goes on.

While it may seem premature to delve into these topics, it won't be long before engineering teams and consultants find themselves engaged in the challenging task of integrating AI systems into the existing fabric of modern corporations and institutions.

The Day After

While it is true that nobody can accurately predict the future, this does not mean we should refrain from proactive thinking and preparation. So, what does the day after look like?? We could find ourselves in a drastically transformed world. Consider, for instance, a scenario where unemployment among white-collar workers surges from 4% to 20%, a fivefold increase. This represents a substantial portion of the workforce, comprising individuals with average to above-average income levels. Consequently, the economic impact would exceed what the mere fivefold increase in unemployment suggests.

The decline in income would be partially offset by highly skilled professionals whose productivity would experience a noticeable boost, leading to increased demand for their services and subsequently higher salaries. However, their numbers are unlikely to compensate for the loss in the lower tiers of the workforce.

Such a situation would place immense strain on social safety nets, such as unemployment benefits, and would inevitably ripple through the economy, leaving many more consumers financially constrained. Consequently, a reevaluation of social benefits may be necessary. Some nations may opt to scale back benefits, while others may choose to increase taxation for those still employed.

Those who benefit from the new order may enjoy additional perks alongside higher salaries, such as extended holidays to indulge in their newfound wealth, unless excessively burdened by taxes.

Unlike the labour shift from fields to factories during the Industrial Revolution, we cannot guarantee that enough new job opportunities will emerge for those adversely affected in this scenario.

Many individuals who are on the verge of losing their jobs will be compelled to reenter the labour market, even if it means taking up more labour-intensive tasks. Those with higher education already possess an advantage over their less educated counterparts, and the possibility of labour displacement should not be ignored.

In less developed countries, where unemployment levels are already high (often obscured by a sizeable informal economy), reliance on agriculture or production may provide a buffer against the immediate impact of technological advancements. For once, technical progress might not make them the primary victims.

Conversely, more developed countries may witness a reversal of fortunes. Higher unemployment rates and a potentially growing population of people without jobs could lead to shifts in migration patterns. This does not imply that Europeans will seek opportunities in Africa or that Americans will flock to Mexico for work. Rather, it suggests that Europe and North America may become less attractive and even less welcoming when their own citizens are not adequately supported as they were in the past.

In summary, we find ourselves at the dawn of yet another revolution. The extent of the damage, whether physical or emotional, will largely depend on the speed of the transformation, much like what occurred during the Industrial Revolution.

What are your thoughts on this topic?