From free convenience to metered intelligence
https://arab.news/96nnj
The evolution of chatbots over the past decade reads less like a steady march of innovation and more like a cautionary tale about how quickly convenience can turn into dependency, and then into cost.
In the early days, conversational AI felt like a small miracle: tools that could write, summarize, translate, and brainstorm at no apparent cost. For many, it seemed like getting “a good job for nothing,” a digital assistant that never slept and rarely complained.
This first phase was defined by abundance. Tech companies, eager to capture market share and the public imagination, heavily subsidized access. The barriers to entry were virtually nonexistent. Students used chatbots to draft essays, professionals used them to automate emails, and businesses experimented with customer service solutions. The implicit message was clear: this was the future, and it was free — or close enough to feel that way.
But, as with many digital revolutions, the honeymoon phase did not last.
The second phase arrived quietly: subscriptions. What was once freely available began to sit behind paywalls. At first, the fees seemed reasonable, even justified. After all, these systems required immense computational power, constant updates, and teams of engineers. Paying a modest monthly fee for enhanced capabilities — faster responses, greater accuracy, and more advanced models — felt like a fair trade.
Yet this shift marked a turning point. The relationship between users and chatbots changed from curiosity-driven exploration to utility-based dependence. People began integrating these tools deeply into their workflows. Writers relied on them for drafts, coders for debugging, and marketers for campaigns. The more indispensable chatbots became, the easier it was to justify the subscription. And the more users justified it, the more entrenched the model became.
Then came the third phase, and with it, a more troubling reality: displacement.
As businesses realized the efficiency gains offered by AI, the calculus began to change. Why hire a junior employee to draft reports when a chatbot could do it instantly? Why maintain a large support team when automated responses could handle the bulk of inquiries? In many sectors, especially those reliant on routine cognitive tasks, jobs began to shrink or disappear altogether.
This was not the dramatic, overnight job apocalypse often predicted in science fiction. Instead, it was gradual and uneven. Some roles were redefined; others were quietly eliminated. For many workers, particularly entry-level professionals, the ladder they expected to climb seemed to vanish. The irony was hard to ignore: the same tools that once felt like empowering assistants were now competing with their users.
And now, we find ourselves in what could be called the fourth phase: metered intelligence.
The pricing models are shifting again, this time toward usage-based billing that can feel, at times, startlingly high. Instead of a predictable monthly subscription, users are increasingly confronted with per-query costs, token limits, and tiered access that resemble utility billing. The more you rely on the system, the more you pay.
For heavy users — businesses, researchers, and even individuals who have woven chatbots into their daily routines — the costs can add up quickly. It is not uncommon to hear comparisons to electricity bills: a necessary expense that fluctuates unpredictably and is difficult to reduce without sacrificing functionality.
This raises uncomfortable ethical questions. If AI tools are becoming essential infrastructure for modern work, should access to them be governed purely by market dynamics? When productivity tools become expensive enough to exclude individuals or small businesses, the risk is not just financial — it is societal. We risk creating a divide between those who can afford augmented intelligence and those who cannot.
Moreover, the economic logic begins to look increasingly paradoxical. Companies may save money by reducing staff and replacing them with AI systems, only to incur significant ongoing costs for using those same systems. For smaller organizations, the equation can become even more strained. What initially appeared to be a cost-saving measure may, over time, rival — or even exceed — the expense of human labor.
In this context, a somewhat ironic conclusion emerges: for certain tasks, it might indeed be cheaper — and perhaps even more sustainable — to hire two interns rather than rely exclusively on a chatbot. Human workers bring not only labor but also judgment, creativity, and a capacity for growth that no billing model can fully capture.
As we move forward, the challenge will be to strike a balance. Can we design economic models that sustain innovation without exploiting dependency? Can we ensure that the benefits of AI are broadly shared rather than concentrated among those who can pay the most? And perhaps, most importantly, can we remember that technology is meant to serve human needs — not redefine them in ways that leave people behind?
The story of chatbots is still being written. Whether it becomes a narrative of empowerment or exclusion will depend not just on what these systems can do, but on how we choose to integrate them into our economic and ethical frameworks.
• Rafael Hernandez de Santiago, viscount of Espes, is a Spanish national residing in Saudi Arabia.
