Do you believe IDC’s report that 72% of 2100 global business leaders report their companies using AI average 3.5x returns in under 12 months?

Wrong question. Here’s where my brain is at…

In 1865 William Stanley Jevons observed the inverse correlation between increasing efficiencies of technology and its consumption (specifically related to England’s coal consumption at the time).

The Jevons Paradox occurs when the efficiency of a technological innovation reduces the resources required yet the falling cost spikes demand that increases its consumption. Said another way, the more efficient something becomes the more we consume.

This counterintuitive outcome occurs because the efficiency improvement lowers the effective cost of using the resource, leading to broader and more intensive use that outweighs the efficiency gains.

When considering adoption and use cases of generative AI for civilization through the lens of Jevons Paradox, five intriguing parallels and implications emerge.

1. Increased AI Efficiency Leads to Greater Consumption
2. Broader AI Access and Democratization
3. Environmental Impact of AI
4. Societal and Ethical Implications of AI
5. Impact on Human Employment and Skill Development

Increased AI Efficiency Leads to Greater Consumption

Generative AI is becoming more efficient and cost-effective for producing content, solving problems, and automating tasks. Ecosystems of interconnected and bespoke applications interact in ways that improve ROI while the raw cost of AI models continue to fall.

This is not a novel occurrence to us in modern times. We are conditioned to this precept through technology advancements already. Yet, we often fail to grasp that the more effective a technology is at delivering value with a lower cost the more we tend to consume.

According to Jevons Paradox, rather than leading to a net reduction in human effort or resource consumption, the increased efficiency could result in even greater consumption of these technologies–and by extension underlying resources. This likely will manifest as continued expansion in the scope and scale of tasks we automate or seek assistance using AI. We are already seeing the dawn of AI-generated content and solutions.

Will this growth slow at some point?

There are four key constraints we have to consider as we peek into our future:

1. Energy – Energy is not ubiquitous nor is it cheap (presently). As there is an emerging arms race in the realm of AI, expect pressure to increase production and lower cost of energy to skyrocket. This may come in the form of new types of energy, loosened sanctions, increased hostilities, etc.
2. Compute – Generative AI consumes massive amounts of compute, which in turns requires energy and processing hardware. Forecast demand for future models is driving headlining discussions about our future in compute production.
3. Data – Access to information to train our models is a critical. LLMs are especially hungry for data. Much like energy, data is not unlimited. Can we generate enough data to ensure new generations of models continue to improve?
4. Humans – Following the Law of the Lid, innovations in AI are unable to surpass our abilities to ideate new approaches to next level AI. At some point, this will reach a tipping point where AI will help us (or help itself) ideate on new innovations we could not see ourselves.

We should expect to see continued acceleration of AI use cases; I don’t believe it will be a continued hockey stick of growth, because encounter a variety of the key constraints manifesting in different ways.

That said, world governments are signaling the advent of an AI arms race. Expect nation-states to focus tremendous effort towards overcoming obstacles as we press against key constraints.

Broader AI Access and Democratization

The increasing efficiency and cost-effectiveness of generative AI could lower barriers to entry, making these technologies accessible to a wider range of users and industries.

Long term benefits of the offering remain relatively expensive and will see increased democratization as price points fall (or the cost is otherwise subsidized).

This democratization could lead to a proliferation of new applications and enrichment of the broader ecosystem, some of which will likely drive up the overall consumption of computational resources and data, in line with Jevons Paradox.

The utopian hope for a new technology is that it improves lives for the better. Much like the weightloss drug Ozempic, GenAI has unarguably crossed the chasm and moved into use by late-stage early adopters. It is arguably on its way to being employed by the early majority, where it will likely encounter increased friction as the reality of what GenAI can and can’t do sets in.

Waves of newer models will rise, crest, and fall upon increasing masses of users as businesses fully implement solution-sets as point solutions or perhaps broader ‘AI operating systems’ (as Sara Davison is coining). This relentless tide will eventually overcome the friction of what it couldn’t do as function finally arrives at (and eventually exceeds) parity with human talents of all manner.

And much like Ozempic is finding its way onto drug formularies with long-term positioning aimed at solving national obesity and metabolic crises, GenAI is painted as a savior to help others reclaim time, be more productive, and focus on what YOU desire.

But it’s not all about you. This is just the medicine you are being fed.

Ozempic is a solution to solve a dietary (and educational) problem largely caused by generations of poor policy, self-interested lobbies, and a host of other social misqueues (including fixation on doing and becoming, rather than being). While I will not deny the outstanding benefit of this class of drugs, it is not lost on me why they are needed.

If and when we all get our own personal assistant, it will likely offer a similar salve to apply to whatever ails us. But will this new fix-all intelligence liberate or further shackle us?

I struggle with democratization of this technology, not because what it can do. Rather, I fear it will amplify our own brokenness. We have to look no farther than our affair with Social Media. Perhaps, it can quietly whisper encouragements while gamifying our personal improvement journey. It is far more likely it will soothe our broken egos with sycophantic syllables–not because of evil design, but because we ask it to.

Environmental Impact of AI

The increasing use of AI could lead to greater energy consumption and carbon emissions, as data centers powering these technologies require substantial energy. The efficiency improvements in AI algorithms and hardware may not suffice to offset the increased consumption driven by their widespread adoption and use.

The paradox provides a framework for understanding the environmental implications of increasingly prevalent AI technologies:

1. Increased Energy Demand from Greater AI Utilization:

Generative AI is becoming an indispensable tool. Improvements in AI efficiency often lead to expanded uses rather than reduced energy consumption. Every new application of AI increases the overall energy demand as more tasks are automated and more processes are optimized through AI.

2. Efficiency Gains vs. Scale of Use:

While advances in AI algorithms and the hardware used to run them have become more energy-efficient, these gains are often outpaced by the scale of AI deployment. As AI becomes more accessible and cost-effective it is reasonable to expect a significant rise in energy use. Increased efficiency in computational tasks does not necessarily translate into reduced energy consumption; instead, it can lead to a broader adoption that magnifies energy use.

3. Data Centers and Carbon Emissions:

The backbone of generative AI is the data centers that train and run these models. These centers consume large amounts of electricity, much of which is still generated from non-renewable sources. As AI use grows, so does the demand for more robust data centers. The energy efficiency improvements in data center technologies are crucial, yet they struggle to keep pace with the rapid expansion of AI services.

4. The Role of Renewable Energy:

There is a pressing need to power AI technologies. Some assert renewable energy sources are key to its future expansion. Transitioning data centers to renewable energy could better align technological advancement with environmental sustainability. However, the availability and integration of renewable energy sources remain key challenges.

5. Systemic Changes and Policy Implications:

Merely improving technological efficiency is insufficient as the sole strategy. Systemic changes, including policy interventions, are necessary to manage the environmental consequences of AI’s expansion. Policies could include regulations to improve the energy efficiency standards of AI systems, incentives for using renewable energy, and measures to offset other constraints to AI operations.

Societal and Ethical Implications of AI

The widespread adoption of generative AI, fueled by the mechanisms outlined in Jevons Paradox, poses societal and ethical questions. Issues such as misinformation, privacy, and socioeconomic division could be exacerbated as the use of these technologies expands.

The sophistication of misinformation and threats continues to grow. We’ve seen the massive effect some misinformation campaigns have had in the US. But many have missed more insidious experiments by actors like Cambridge Analytica corrupting election processes in countries such as Trinidad using Facebook.

Have you taken a look at your credit and identity monitoring reports lately?
A quick tour through any of these for-pay services offers ample evidence of the  volume of privacy violations individuals endure every single day. Google just updated its terms of service to include AI ToS as standard fair, signaling an important shift in how companies intend to harvest, use, and leverage our personal data.

Who do you trust?

Where does it stop? What is the balance between personalized, context-sensitive experience and abuse? Who gets to control that decision?

More importantly, the advent of AI in the coming years will exacerbate the divide between the haves and have-nots. While telephony and internet access are now arguably part of Maslow’s Hierarchy, access to AI is far from ubiquitous.

The purveyors of such platforms advertise laudable changes to our socioeconomic conditions, where the new currency of information will power a new economic paradigm (complete with cheap, abundant energy). I’m not really sure how we are expected to believe this when executive and elected official alike are thin on “how to” answers.

Ah, the age ole’ capitalist mantra “the market will decide.”

No, a future with AI requires careful consideration and balance in building bridges between our present and our future if we are to survive and thrive. (It’s the thriving part I question, frankly.) We require policy crafted by lawmakers more informed about what Wi-Fi actually is and business leaders more in tune with stakeholders than shareholders.

Alas, the paradox would have us recognize a different reality–a reality where appetite drives innovation in hedonistic directions, tempered meagerly by lagging policy until a tipping point of utter frustration is eventually reached. Leaders will be wise to heed history’s lessons that great disparity leads to outright revolution.

Impact of AI on Human Employment & Skill Development

Generative AI: Job destroyer or bringer of economic freedom?

The narrative surrounding artificial intelligence often vacillates between dire predictions of widespread job displacement and utopian visions of limitless economic opportunity. AI becomes more efficient, the scope of its application widens, and rather than constricting employment opportunities new opportunities emerge.

AI, by automating routine tasks, may initially diminish the demand for certain skillsets, seemingly validating fears of it as a ‘job destroyer’. However, the broader perspective reveals a more nuanced tableau. This technological evolution could foster a surge in demand for new skills and roles, potentially catalyzing an era of unprecedented professional transformation.

This dual-edged sword cuts through the labor market, severing outdated roles while simultaneously stitching new ones into the fabric of the economy. For example, the rise of AI-driven analytics tools demands a higher level of data literacy across all sectors, thereby expanding the labor pool to include roles such as AI trainers, overseers, and maintenance specialists.

Moreover, the democratization of AI tools could lower barriers to entry for entrepreneurs and innovators, propelling a wave of start-up culture focused on leveraging AI for bespoke solutions. This shift towards an ‘AI-first’ business model could redistribute economic power, moving it from the hands of the few into the hands of the many.

Could we see a 1-10 employee “unicorn”?

Yet, let us not naively embrace this as an unalloyed good. The infusion of AI across industries carries the potential to exacerbate existing inequalities unless carefully managed. The critical question we face is not whether AI will replace jobs, but how we can mold its impact to ensure it serves as a tool for economic empowerment rather than disenfranchisement.

As we navigate this transformative period, the real challenge lies in our ability to adapt our educational systems and policy frameworks to keep pace with the rapid evolution of AI capabilities. Only through thoughtful, inclusive dialogue and proactive strategy will we harness the paradoxical nature of AI—turning potential job destruction into a beacon of economic freedom and skill development.

In the end, it’s not the machines that decide our fate, but how (and how quickly) we choose to integrate them into the very sinew of our societal framework.

A Journey In Paradox

How does a 19th-century economic paradigm relate to GenAI? We’ve explored five intriguing parallels through the lens of Jevons Paradox.

It’s evident that the paradoxes presented by generative AI are not mere academic curiosities but pressing realities that will shape the future of our global society.

GenAI is not a standalone revolution but an evolution that intersects energy, data, policy, and human ingenuity deeply. The increased efficiency of AI does not simply promise reduced costs and enhanced capabilities; it heralds a potential shift in how we interact with technology, each other, and our planet.

1. Revisiting Consumption: As AI becomes more efficient, its applications expand, leading to greater consumption of resources. This increased demand challenges us to find sustainable ways to fuel this growth without depleting our resources.
2. Democratization and Access: While AI promises to lower barriers and democratize access, it also poses the risk of amplifying disparities if not managed with foresight and equity. Ensuring broad and fair access is paramount to harnessing AI for societal benefit.
3. Environmental Stewardship: The environmental impact of AI is profound. As stewards of our environment, we must advocate for and implement AI solutions that prioritize sustainability, leveraging advancements to reduce, not increase, our demands upon our environment.
4. Ethical Considerations: The societal and ethical implications of AI are perhaps the most complex to navigate. From privacy concerns to misinformation, the way we address these issues will define the ethical landscape of our future society.
5. Employment and Skills: Finally, the impact of AI on employment and skill development is double-edged. While AI displaces some jobs, the paradox biases us to believe new opportunities for those who adapt. The challenge lies in preparing our workforce for these inevitable changes.

While Jevons Paradox provides a valuable framework for understanding the potential impacts of AI, it also compels us to think critically about how we can use these insights to shape a future that leverages AI’s power responsibly and inclusively.

This journey has not been about finding definitive answers. Rather I hope to spark conversations that prepares us to navigate the complexities of a world intertwined with AI. As we move forward, let us not be daunted by the scale of these challenges because they will happen. We must find a way to balance inspiration at the potential to craft a future that reflects our highest aspirations and deepest values against the recognition to which our biases blind us to our own selfish intents.