Don’t have time to read the full article? This summary covers its core concepts in 300 words. Use the links to navigate to any section of the article that interests you.

• Understanding AI:
  • Artificial Intelligence Defined: Distinguishing between current narrow AI capabilities and the speculative potential of artificial general intelligence (AGI).
  • Current Limitations: Explanation of AI’s current role as powerful but limited tools that do not truly “understand” or “think”.
• Mechanics of AI Tools:
  • Generative AI: How tools like ChatGPT use pattern recognition based on vast datasets to generate responses.
  • Application Specifics: Explanation of AI’s domain-specific applications, such as language processing versus self-driving technology.
• AI in Current B2B Marketing:
  • Operational Efficiency: Examples of AI in streamlining research, content generation, and data analysis.
  • Creative Applications: Use of AI in generating marketing materials and managing customer interactions.
• Ethical Considerations of AI usage:
  • Intellectual Property: Concerns over ownership and profits from AI-generated content.
  • Data Privacy and Usage: Ethical implications of using large datasets, potentially without consent.
  • Long-term Impacts: Potential for job displacement and societal changes due to AI advancements.
• Historical Parallels:
  • The Printing Press: How it democratized information and standardized communication, mirroring potential AI impacts on information accessibility.
  • The Industrial Revolution:
    • Productivity and Innovation: A feedback loop of innovative technology leading to increased profits leading to increased investment in new technologies revolutionized the global economy.
    • Societal Shifts: Emergence of capitalism and liberalism, the long fight for worker’s rights.
  • The Age of the Automobile:
    • Economic and Social Transformation: Shift from horse-based to machine-powered transport and its implications on workers, urban planning, and industries.
  • The Digital Age
    • Emergence of a purely digital economy that interacts with the physical world
    • Invention of entirely new industries that would have been unimaginable mere decades earlier.
• Predicting AI’s Future:
  • Integration and Automation: Speculation on AI’s role in creating unified systems for sales, operations, and marketing.
  • Market and Employment Changes: Predictions on how AI could redefine job roles and business strategies.
• Potential Societal Impact:
  • Work Redefinition: Discussion on the possible reduction of the workweek and shifts in workforce requirements.
  • Economic Disparities: How AI might widen or narrow economic gaps depending on policy and implementation.
  • Cultural and Social Adaptations: Changes in education, employment, and possibly social values due to pervasive AI integration.

 

They say there are two inevitabilities in life: death and taxes. Let’s add a third—progress.

Progress, be it political, social, or technological, is unavoidable. Progress isn’t always linear—humanity can and has taken steps backward through history—but on a large enough scale, it’s proven inevitable.

With each major societal and technological milestone comes sweeping and sometimes unexpected change. Consider the horse trade during the early 20th century, as the automobile began to replace the horse-drawn carriage. The obvious impact would be fewer horses, fewer stables, less work for farriers and saddle makers. But what about the impact on agriculture? Demand for hay plummets at the same time as gas-driven tractors emerge, greatly improving productive output. How did this knock-on effect shape the decades to come?

We are in the midst of another transformative technological event: the emergence of generative AI. We’ve discussed many ways that AI has impacted B2B marketing, from SEO to market research and web development. Today, let’s take a big picture look at AI, its current and future roles in sales, operations, and marketing, and the possible impacts it will have on other areas of the economy.

What is AI? Defining AI in the context of B2B business operations

There can be some confusion when talking about what we broadly refer to as AI. We know that AI means “artificial intelligence,” but we also know that today’s AI isn’t quite intelligent. A tool like ChatGPT doesn’t understand the answers it provides. Under the hood it’s like an extremely powerful autocomplete, regurgitating information without any true analysis.

The science fiction version of AI is commonly referred to as “AGI” or “Artificial General Intelligence.” The tools we have now are both “unintelligent,” meaning they lack any understanding or sense of self, and “domain specific,” meaning they cannot perform adequately outside of a narrow range of applications.

ChatGPT and other tools can answer questions on many topics. They cannot drive cars. Emerging self-driving technology in vehicles can drive cars, but it cannot answer questions. This is domain specificity in practice.

One day—perhaps soon, but likely not for many years—we may achieve AGI. A new type of AI that understands itself in relation to others, is capable of thought, and can accomplish many types of tasks rather than a select few. It could even feel.

If that feels a bit too Skynet for your liking, you aren’t alone. Futurists and neo-Luddites alike speculate on the dystopian possibilities of AGI. For now, let’s set aside fears of our future robot overlords and focus on the kind of AI we have now—which is “AI” in name only.

How do generative AI tools work?

Let’s dig a bit deeper into how a tool like ChatGPT actually works.

ChatGPT is built upon large language model (LLM). It is a chat interface that queries this model to generate human-readable responses to any number of informational queries, ranging from Wikipedia-like responses to historical questions to generation of working code in a variety of programming languages.

Let’s look at how an LLM works. Large Language Models start with a corpus—a massive volume of text formatted to be read by a computer. In the case of ChatGPT, enormous volumes of text from across the internet including news articles, blog posts, and technical documents was used.

The Large Language Model digests this massive volume of information, breaking it down into patterns and sequences. When we ask ChatGPT questions, its model spits out content with the correct pattern or sequence to answer the question. The model has been tuned by its engineers to correctly parse and respond to colloquial prompts.

Let’s imagine we ask ChatGPT a question like “how does an internal combustion engine work?”

Its model will have “read” dozens or hundreds of answers to this question from around the internet. The generated answer will be something like an average of the answers it’s read, adjusted for any context provided. For example, it will answer differently if you ask for an explanation suited to a child than it would if asked for an explanation suited to an engineering student.

It is critical to understand that at no point does ChatGPT think about the question or verify its reply. If the model contains false information, ChatGPT may deliver that false information.

Additionally, tools like ChatGPT have invisible internal prompts that guide their responses. An unbridled LLM could respond to any query, providing dangerous or offensive answers. To mitigate that risk, internal prompts will forbid the model from answering in certain ways, using certain language, or even replying at all to prompts that may violate terms of service or pose ethical concerns.

Other forms of generative AI

Today’s AI models go far beyond text, with generative tools available for every form of media from images to music and video.

Regardless of the type of generative AI, the principle of operation is similar. A model digests vast amounts of media. Its engineers guide it to produce useful outcomes and implement safeguards to prevent abuse. It generates responses based on the data it’s digested, and the prompt given, with no thought or consideration for what’s produced—except whatever rules it’s been programmed to follow.

AI’s role in B2B marketing today

Generative AI is a product in an emerging AI industry. It’s sold to marketers as a means to simplify their day-to-day tasks, speed up production, and reduce hours spent on menial work.

These benefits have been realized by agencies around the globe. Many agencies go even further, tasking AI directly with content creation and other “creative” tasks. Our position is that it’s a bad idea to make direct use of AI for website content or advertising assets—but companies do it, with mixed degrees of success.

Here are a few ways we actively employ ChatGPT in our day-to-day tasks at BlackBean:

• Verification and expansion upon document outlines for blog posts and articles
• Summarization of lengthy and complex documents for easier reading and interpretation
• Initial planning and brainstorming exercises for research projects
• Generation of non-critical image assets for newsletters and other content
• Testing and production of scripts and code snippets for web and desktop projects
• Cleaning and processing messy or incomplete marketing data (directly or through generated scripts)

Additionally, many popular marketing tools incorporate AI directly either as part of the user experience or behind-the scenes. We actively or passively use AI features in tools such as:

• Google Ads
• Google Analytics
• Notion (project management tool)
• Search engines like Google and Bing
• Landing page builders like Leadpages

As you can see, today’s marketing world is already saturated with AI. While the high value stuff like research, strategy, and campaign execution remains in the hands of skilled humans, many less pressing tasks can be accomplished quickly with AI assistance.

Example of an AI-assisted marketing workflow

Let’s look at a common use case for AI marketing assistance: planning and launching a search advertising campaign.

Appropriate planning for a search campaign—especially in the often expensive and complex realm of B2B lead generation—is a data-intensive process. An advertiser will want to know as much as they can about their competition, expected costs, ideal target keywords, customer personas, and more.

AI can assist with the collection and high-level of analysis of each of these factors, allowing a human mind to synthesize the prepared information into an effective advertising strategy. Let’s explore a typical process:

1. The advertiser begins with keyword research. Using SEM research tools, they analyze top search keywords for purchasing intent, competition, and volume. After identifying competitors, the advertiser may crawl competing websites to verify findings, using AI to parse the website content into top keywords and topic clusters.
2. After loosely organizing keyword data into a master spreadsheet, the advertiser feeds it into an AI program with instructions to identify themes and organize the keywords accordingly. After reading the output, they make manual adjustments as needed.
3. The advertiser writes drafts for the headlines and descriptions of each ad group’s primary search ad. They use AI to brainstorm variants based on the initial drafts, greatly expanding the number of advertising assets created in a short period of time.
4. Once the campaign launches, a different kind of AI continues to be involved—passively and behind the scenes. Google relies on its own machine learning technology to identify best-fit users for each ad, gradually improving performance over time based on conversion and demographics data.
5. As the advertiser continues to manage and optimize the campaign, they may use AI to assist with reporting—generating attractive graphs from raw data within seconds that might otherwise take minutes or even hours. They may also use AI to identify abstract patterns in the data that a human analyst may miss. These patterns can help them further optimize the campaign.

Throughout this process, the advertiser uses their expert insight to determine the correct course of action. AI is a passenger in this process! Savvy use of AI in marketing requires an understanding of the ethics of AI usage as well as the ability to curate AI output, improving upon quality results while rejecting poor content and unhelpful analysis.

Ethical AI usage in B2B applications

The ethics of generative AI are complex. Debate around appropriate usage is ongoing. Expect that debate to intensify in the coming years as AI produced media becomes more and more difficult to discern from human creations.

The issues concerning ethical AI usage stem from a few sources. First, many ethicists wonder about media rights. If a person generates artwork using AI, do they own it? Does the platform that generated the art own it? Does it exist as some sort of open source, rights-free creation for public use?

Others argue that training data used to build these AI models runs afoul of intellectual property rights. If an AI model gets trained on data from social media platforms like Reddit, does the model owe Reddit a cut of its revenue? If the developer scrapes content from millions of personal and professional blogs, why don’t the content creators get any kickback or recognition?

Finally, some ethicists oppose AI in general terms, citing unknowable risks and potentially catastrophic economic implications as the tech advances. These fears are grounded in reality. So-called “deep fakes”—hyper-realistic AI generated renderings of real people in often compromising or misleading situations—have already become commonplace. They are just one example of how AI can enable powerful forms of propaganda and misinformation.

When considering AI ethics in your own business, focus on fairness and avoid introducing risk. For instance, don’t allow an AI chatbot to handle sensitive customer interactions where a mistake might mean financial or legal consequences for your client or business. Be mindful of the shifting legislative landscape regarding intellectual property rights. Don’t use AI to “rip off” other businesses or human creators.

The age of (mis)information: AI’s role in eroding truth

AI’s ability to warp reality isn’t limited to deep fake videos. An interesting attribute of a thoughtless content generator is its capacity to lie convincingly. An unrestrained AI model doesn’t care about the truth. It generates content that most effectively addresses the prompt it’s given. Unrestricted AI could generate thousands of misleading, biased, and dangerous “news” articles a day, engineered to appeal to and persuade specific demographics into believing falsehoods.

Criminals can use AI to accurately mimic voices. Doctoring photos or videos is easier than ever. Detecting AI content is possible, but as tools become more sophisticated—and as more of what we see online is indeed AI generated—it will become increasingly difficult to tell what’s real and what isn’t.

This is the most immediate ethical concern for generative AI. The downside of our internet-connected world is the speed at which misinformation can proliferate. AI makes the challenge of knowing what’s true even more difficult.

Where are we going? The AI Evolution in historical context

Now that we’ve covered some fundamentals of generative AI and its uses in the business world, let’s try to understand where it’s headed next.

If we accept that AI has the potential to be as transformative to businesses as the internal combustion engine or the personal computer in decades past, we may look to those past innovations for context. Let’s explore the outcomes of some major moments of progress, from the printing press to the digital age.

The printing press: enabler of mass media

The invention of the movable type printing press by German inventor Johannes Gutenberg in the fifteenth century ushered in an era of intellectual exploration throughout the western world.

Though similar printing presses existed in China and Korea hundreds of years earlier, Gutenberg’s version, which used the simpler German alphabet rather than the complex series of characters used in East Asian languages, proved highly efficient and effective.

The primary benefit of the printing press to society was the democratization of knowledge. Prior to its invention, even a simple book would take months to scribe. Important tomes, particularly religious works where elaborate illustrations were common, could take several years to complete. To own a single book—let alone to know how to read it—would mark a person as wealthy and sophisticated.

Although mass printed works and hand-scribed texts overlapped for decades or even centuries after the printing press’ invention, the press spelled the eventual end of traditional scribes. A critical role in society for thousands of years across all literate cultures, the role of the scribe turned from artisan to technician as those involved in traditional bookmaking often became press operators and editors.

In the context of AI’s impact on society and the modern economy, it’s interesting to consider the way the printing press transformed monastic life. Monasteries served as centres of religious and intellectual life in Europe for centuries. As the printing press supplanted scribes, it’s possible that it helped refocus monastic life on religious practice. Scribing was never a primary concern for monks in medieval and Renaissance Europe. Simplifying this laborious task while increasing access to written materials may have allowed monks to focus on matters spiritual—perhaps contributing to rapid religious and philosophical transformation during this time period. This is the not-yet realized promise of generative AI: to remove boring but laborious work and replace it with high value, thoughtful activities.

Once the printing press enabled mass production of books and other documents, the educational paradigm began to shift. Easier access to the bible allowed for the emergence of Protestantism. Its focus on the bible and the cultivation of a personal relationship with God through its text could only be made possible by wide-spread access to the bible. Religious groups associated with Protestantism helped usher in educational reforms that gradually improved literacy throughout Europe.

As literacy increased, language evolved. Spelling and grammar became standardized. In fact, the printing press helped standardize many things. Prior to widespread literacy, many industries relied on information passed down via word-of-mouth from generation to generation. This meant that major disruptions to a local population could mean the loss of crucial industry knowledge.

Once a sizeable share of the general population could read and write, seemingly mundane details like recipes and building techniques could be recorded. Scholarship expanded from matters of theology, philosophy, mathematics, and science to practical matters like baking, building, and farming.

One final innovation directly resulting from the printing press is the creation of mass media. Imagine living in a world not just pre-internet, but pre-television, pre-radio, and pre-newspaper. A war breaks out in Germany. How do you learn about it? How long after the war commences does an average person in London know it’s happening? Before the printing press, the world was small. News spread via rumours from travelling merchants, or from letters between diplomats disseminated—often as gossip—to courtiers. The invention of mass-produced newspapers allowed news to spread to common people within days or weeks rather than months or years, as may have been the case prior to widespread literacy.

The Industrial Revolution

The Industrial Revolution is itself an indirect outcome of the printing press. Although centuries separate the printing press from the early days of the revolution, the conditions that allowed for the revolution would not be possible without an educated and innovative mercantile class. From advancements in political and economic theory to technological innovations resulting from greater scientific literacy, the industrial revolution is also an information revolution.

Historians pinpoint the beginning of the Industrial Revolution around the mid eighteenth century in Britain. The introduction of more efficient machinery in textile production resulted in explosive productivity growth, which in turn funded expansion of textile factories. A core concept of the industrial revolution is this feedback loop: technological innovations greatly increase productivity, resulting in lower production costs, resulting in higher profit margins, resulting in more investment in technological innovation.

 

What started in textiles soon expanded to many industries. Iron production became cheaper and more efficient, which in turn allowed railroads to expand. Speaking of trains, steam power transformed many industries, removing the need for hard manual labour from a number of processes.

The industrial revolution continued for over a century, its innovations completely transforming the economic and demographic profile of western society. Prior to the revolution, the west was largely agrarian. Low productivity resulted in slow economic growth. As the effects of the revolution unspooled, more and more people moved to cities seeking economic opportunities. Rapid technological progress in agriculture, including mechanical solutions like the seed drill and mechanical reaper, contributed to far greater crop yields while requiring fewer farmers.

The sociological impact of the industrial revolution was initially negative for many people. Transitioning from mercantilism to early free market capitalism resulted in horrific working conditions for early factory workers. Increased productivity contributed to increased expectations of productivity, which resulted in brutally long workdays, an increase in child labour, and troubling exploitation of subjugated peoples in colonies around the world.

At the same time, capitalism fostered a form of liberty practically unheard-of in prior eras. Ideas of social and economic mobility and entrepreneurship resulted in a liberated middle class—at least when compared to the serfdom and slavery of previous generations. After the industrial revolution, a person could conceivably rise from the lowest ranks of society to the very highest. Consider the emergence of John D. Rockefeller, eventual richest man in America, from his humble beginnings. Born in the latter years of the industrial revolution to a con-man father and a homesteading mother, Rockefeller would go on to become a titan of industry. A century earlier, he may have toiled away his life in obscurity as a poor farm worker.

Liberty came slowly to the working poor in cities across the western world. It was nearly a century after the revolution began before the United Kingdom introduced its first labour laws targeted at factory workers—restricting working hours for children under thirteen to nine hours per day. It would take until the early 20th century before the abolition of child labour and the introduction of minimum wages. Industry eventually paved the way to health, wealth, and prosperity for countless millions, but the path to modernity would be a difficult one.

When examining the industrial revolution’s consequences on society, we see the way that new ideas and technologies can usher in a golden age of commerce and social advancement. We also see the potential cost of such a transition. Before leading us into the modern age with all its benefits, the industrial revolution brought untold suffering, poverty, illness, and death to the working class and to colonized peoples around the world. It introduced new ecological challenges—such as carbon emissions from burning fossil fuels—that humanity still needs to reckon with.

The age of the automobile

In 1900, major cities like New York contained tens or even hundreds of thousands of horses. Where today we debate environmental issues like air quality and carbon emissions, debates at the time centered on manure. Each horse could produce up to thirty pounds of manure every day. Multiplied by many thousands of horses per city, this meant that streets were effectively paved in manure. Cities were built to accommodate horses. Stables existed where we might expect parking lots and automotive shops to be today. Young men might find work as farriers, ostlers, wheelwrights, stablemasters, and coachmen.

 

Less than twenty years later, New York would retire its last horse-driven streetcar. Young men looked for work as car mechanics, service station technicians, and professional drivers. By 1930, there were over 20 million registered motor vehicles in the United States.

Meanwhile, the number of registered horses in the United States peaked at around 19.5 million in 1915. By the 1960s, there were less than five million registered horses nationwide—and over 60 million cars.

The age of the automobile ushered in social and economic changes that are far-reaching, helping shape the world we live in today. Up to 70% of extracted oil and gas gets converted to fuel for vehicles—either gasoline or diesel. Personal and commercial vehicle traffic accounts for up to 25% of carbon emissions each year. The sector employs tens of millions of people worldwide, producing up to $3 trillion in economic output each year.

Big ideas and resistance to change

The automobile age presents an interesting example of resistance to change. In an article first published in 1930, early automobile pioneer Alexander Winton wrote of his experiences trying to build and sell cars in the final years of the nineteenth century:

To advocate replacing the horse, which had served man through centuries, marked one as an imbecile. Things are very different today. But in the ’90s, even though I had a successful bicycle business, and was building my first car in the privacy of the cellar in my home, I began to be pointed out as “the fool who is fiddling with a buggy that will run without being hitched to a horse.” My banker called on me to say: “Winton, I am disappointed in you.”

Later in the article, he quotes a critique levied by a reporter against the very idea of a motorized bus line connecting Chicago and St. Louis:

[…] The fool who hatched out this latest motor canard was conscience-stricken enough to add that the whole matter was still in an exceedingly hazy state. But, if it ever emerges from the nebulous state, it will be in a world where natural laws are all turned topsy-turvy, and time and space are no more.

To citizens of the early 20th century, the world ran on horses. The notion of replacing horses with some sort of machine was, to most people, absurd. It’s easy to look back and laugh at the ignorance of the doubters and nay-sayers of the time, but there were good reasons to doubt the new technology. On one of Winton’s first vehicular voyages, he had to stop at pharmacies along his route hoping they’d have a gallon of gasoline in reserve. There were, of course, no gas stations.

Nor were there parts stores; paved roads; driving schools—or even traffic laws; trained mechanics, or any of the other vehicular infrastructure we take for granted today. There were hundreds of inventors working on strange and radical vehicles, almost all of which failed.

Yet, the naysayers were wrong. Most of them would be proven wrong in their lifetime—in the next decade of their life, in fact. Perhaps the benefit of the automobile was so significant that even the seemingly infinite obstacles preventing its adoption could not stand in the way of progress. In any case, it’s an intriguing example of the power of human innovation. That which we cannot imagine today may be commonplace ten years from now.

The robotic age: modern manufacturing techniques and the rust belt economy

There are many reasons for the emergence of an American golden age in manufacturing after the second World War. Diminished industrial capacity in rebuilding Europe and Asia made American manufacturing crucial and competitive. America’s strong primary sector industry, including vast coal and metal reserves, allowed the manufacturing sector to rapidly expand. The influx of wealth from this thriving economic sector helped it grow even more, as a more affluent middle class spent their newfound wealth on things like cars and home appliances.

However, no golden age lasts forever. As the rest of the world recovered from the devastation of war, international manufacturing became more robust and competitive. Emerging technology, particularly more advanced computer systems, allowed for radical advancements in manufacturing processes. The first industrial robot appeared in a General Motors factory in 1961. The influence of robotics in manufacturing would steadily increase through the coming decades. In 1990, half as many Americans worked in US auto manufacturing facilities as in 1950. Despite this halving of the workforce, the American auto sector produced 33% more vehicles per year in the 1990 as it did in 1950.

Job losses from automation weren’t evenly distributed across the country. Hundreds of thousands of well-paying union jobs evaporated in the American manufacturing heartland—now commonly referred to as the rust belt. This once-thriving Midwest region stretching from Illinois south to Kentucky and east through New York saw its population decline by over four million residents from the mid 20th century to the year 2000. This represented a 35% population decline in the region’s top manufacturing cities.

Despite automation’s clear role in the decline of the rust belt region, the factors contributing to manufacturing decline are varied. High interest rates and appreciation of the US dollar through the 1970s made US exports far less enticing to foreign markets. This economic squeeze on manufacturing both necessitated change—automation could help make American factories competitive again—and depressed employment in the sector.

Another element in this story is trade unions. Automotive manufacturing jobs were famous—in some cases, infamous—for their high pay and strong union support. In some cases, getting fired from a manufacturing job would be nearly impossible, provided the worker showed up for their shift. This resulted in an unusual paradigm for auto workers. Often unskilled, factory workers could raise a family on their manufacturing salary—but couldn’t find any remotely equivalent compensation at other jobs they were qualified for. In a podcast episode of This American Life focused on American/Japanese cooperation in the auto sector, engineering professer Jeffrey Liker describes the “prison-like” scenario that auto workers experienced:

Actually, the analogy to prison is a good analogy, because the workers were stuck there because they could not find anything close to that level of job and pay and benefits at their level of education and skill. So they were trapped there. And they also felt like we have a job for life, and the union will always protect us. So we’re stuck here, and it’s long term.

The gilded cage many workers found themselves in would lead to animosity as automation and economic factors slowly starved the auto sector. It is interesting to consider the popular narrative of declining manufacturing work in the US: robots showed up and took all the jobs. In reality, complex economic and geopolitical factors necessitated the change. Perhaps it’s human nature to fixate on the tangible rather than the nebulous. A robotic arm on an assembly line killing ten factory jobs is easy to grasp. Purchasing power parity and interest rates are more slippery.

The digital age: personal computers, the internet, and social media

Computing enabled the sort of factory automation that radically impacted the manufacturing sector in the United States, but it was just a taste of what was to come. The popularization of the home computer through the 1980s and ‘90s resulted in radical shifts in culture and economics on par with the printing press and the industrial revolution.

Have you ever typed a letter on a typewriter? If so, you’ll understand just how revolutionary a personal computer could be. Even the most rudimentary word processor offers an editing experience unrivaled by any analog system. The backspace key alone is revolutionary.

The printing press enabled mass media for the first time in history. The internet enabled instant mass media for the first time. A news report that may have taken weeks to spread through Renaissance Europe, or days to spread around the globe in the mid-20th century could now be shared across the world in seconds.

This rapid access to information has proven to be a double-edged sword. While information and education have never been more democratized, consumer trust has never been lower than today. Internet users are confronted by spam, scams, and misinformation in a way that never occurred before the digital age. This erosion of trust has resulted in significant transformation in sales and marketing. The average person is far more media-literate today than they would have been in the ‘80s. Where ads of yesteryear often relied on product benefits and simple sales propositions, today’s advertising industry leverages more nuanced consumer psychology to produce results.

If the radical economic transformation brought on by motor vehicles shocked people of the time, imagine trying to explain the way the internet reshaped the economy to people in 1990, let alone 1900. Terms like “app developer”, “ethical hacker”, or “cloud systems engineer” would be meaningless. The estimated economic contribution of the digital economy today is up to 15% of global economic output. Thirty years ago, the digital economy did not exist.

Much like the emergence of liberalism and free market economics during the industrial revolution led to greater social and economic mobility for average people, the emergence of a digital economy has allowed startups and small businesses to ascend to industry titans in a matter of years. Consider tech giants like Google and Meta—founded by college kids in dorms and garages.

The impact of the digital economy reaches into the physical world. The 2000s saw a major increase in online shopping, impacting the bottom line for conventional retailers like Sears. This would be exacerbated by the release of the first iPhone and the smartphone revolution that followed. Within a decade, e-commerce devastated brick-and-mortar retail, typically offering better convenience and better prices. In the ‘90s, Sears operated as many as 3,500 locations across the United States. Today, there are only ten Sears locations still in operation.

Broadly speaking, the digital age marks the beginning of a robust “ideas economy.” Workers that once fed their families by producing goods and providing physical services now earn their wage working on purely digital products and services—things that exist only as bits and bytes on servers.

The future of generative AI

Generative AI and other technologies emerging from AI research and development have the potential to impact society at the same scale as the transition from horses to cars, from analog to digital.

When evaluating what may come next for AI, we can learn from our historical examples. For instance, an early automobile patent by one Uriah Smith described a vehicle with a fake horse head, intended to ensure it didn’t spook the horses it would share roads with.

 

Looking back, we see a flaw in Mr. Smith’s thinking. It took just twenty years from the introduction of the motorcar for travel by horse to become obsolete. There would be no need to avoid spooking horses on the road, because the horses would no longer be on the road at all.

So it is with AI. We should be careful not to make predictions that heavily weight today’s technological paradigm. It’s likely to shift considerably.

We can also be mindful of the complexity of technological transformation. Like robotic arms in factories and the complex economic systems that drove their adoption, AI’s future will depend on many external factors.

We should be mindful of the role of the worker in the future of AI—much like the industrial revolution, things may get worse for workers before they get better. We must be prepared to explore alien ideas, speculate on industries that don’t exist yet. When dealing with incredibly disruptive tech, the only certainty we can expect is disruption.

With that said, here are some possible future impacts of generative AI that we expect to see over the next 25 years, and their potential effect on B2B businesses.

Total systems AI: unified operations, marketing, and sales

While true artificial intelligence may or may not be on the horizon in the coming decades, broader domain scopes for generative AI are coming. Currently, there are AI chat bots that can handle onboarding and basic sales tasks for new customers. There are AI tools that assist with marketing in numerous ways. Some AI tools can analyze operational data and recommend targeted improvements to operations efficiency.

Tomorrow’s AI tools will be able to mesh each of these business operations together into a unified system. Imagine a digital “brain” at the heart of your business, able to react to events in each department. A large order through the sales department would automatically result in a purchase order for supplies required to fulfill the order. An increase in marketing leads from a particular sales region would result in re-allocation of sales resources to help close those deals.

The real transformative aspect of broadened domain scopes in AI would come into focus when many or all businesses adopted this sort of broad business automation. Imagine a world in which your sales AI talks to a client’s purchasing AI, hashes out an optimal deal for both parties, and sends the work order back to your facility—without any human involvement at all.

It might sound crazy—any good AI prediction should! As AI systems become more powerful and reliable, we could see a completely different sales model emerge in the B2B world. Human salespeople and marketers would build personal relationships with companies as always. After two companies decide to work together, they would integrate each other into their sales and purchasing systems, leaving AI to deal with the rest of the process.

The prospect is a bit scary, but intriguing. Think of sales deals that fall through due to human error or even personal incompatibility. A computer is better-equipped to make unbiased business decisions—if we can get to the point that we trust them enough to “hand over the keys.”

Ultra-personalized digital experiences

If we accept that generative AI is something of a pattern-recognition engine, we can see how it could transform marketing experiences. AI can already speed up data collection and analysis. Tomorrow’s AI may be able to react to digital signals in real time, computing the optimal way to reach every single user on your website or app.

Imagine a website that looks different to every single user. While it retains your company’s brand, its layout and design could vary widely. Each headline and sentence on the website would be re-written at the moment the page loads to best communicate your message to every user.

If an elderly bespectacled college professor visits the website, it may shift to become simpler in its design, increase the size of text on the page for readability, and re-write your sales copy to be more technical and verbose. If a twenty-year-old mechanic visits the same page, it may offer a more vibrant and interactive experience using exciting and dynamic sales copy.

This assumes that websites as they exist now will continue to be a primary driver for business. It’s possible that more and more digital experiences happen within an AI-powered system completely unfamiliar to us today. In any case, we predict ultra-personalized marketing experiences are the future.

Re-imagining the work week

Recall the scope of societal change brought on by the industrial revolution. Prior to the revolution, most people lived agrarian lives. In the midst of the revolution, many people—children included—worked extremely long hours in filthy and dangerous conditions. In the decades after the revolution, we saw establishment of minimum wages and standardized work weeks.

AI has already made an impact on employment. Large tech companies have seen the benefits of AI “copilots” for computer programming, streamlining their development departments and laying off thousands of tech workers. As AI grows more capable, it will displace workers in more and more fields.

When workers get displaced by new technology, many of them train to work with that new technology. Scribes become printers. Farriers become auto mechanics. However, new technology results in greater efficiency that, by design, necessitates fewer workers. If AI kills ten marketing jobs and creates 10 equally intensive AI management jobs, it must provide exceptionally greater marketing performance than the original team in order for it to deliver its promised efficiency.

Consider the rise of automation in manufacturing. We know that many workers retrained to work as technicians on the machines that “took their jobs.” But we also see that total workers in American auto manufacturing declined by half from 1950 to 1990. Automation doesn’t create as many or more jobs than it removes.

In that sense, truly disruptive AI technology will necessitate a reimagining of the work week. Did you know that the famous economist John Maynard Keynes predicted in a 1930 article that by the turn of the century, the work week would be just 15 hours long?

He was proven wrong. New kinds of industry emerged that he could not have imagined. The same may be true for AI. If not, it’s possible that his prediction will come to pass, just a few decades late.

Much like the monks whose scribing duties interfered with their spiritual development, we could imagine a world in which less of our time gets spent on operational tasks. Workers may be able to focus on high value and creative pursuits, using more of their free time to benefit their health, well-being, and personal development.

It can sound crazy—utopian, even. But there’s a real possibility, especially with advancements in robotics, that AI will result in shorter work weeks and vastly different expectations in the working world—and sooner than you might think.