A technology shift unlike previous waves
At Software Planet Group, we have spent years helping businesses build, modernise, and scale software products. We have seen many major technology shifts, from cloud computing and mobile platforms to DevOps, automation, and low-code platforms. Each of them changed the way companies designed, delivered, and maintained software. Yet none of them compares to the scale and speed of what is happening with artificial intelligence today.
For many years, AI appeared to be progressing more slowly than expected. Neural networks had existed for decades, but most practical applications remained limited to relatively narrow tasks such as image recognition, classification, pattern detection, and basic automation. The technology was useful, but it did not feel like a general-purpose intellectual engine.
That changed with the emergence of transformer architectures, attention mechanisms, larger training datasets, and increasingly powerful computing infrastructure. AI moved from solving isolated problems to performing a much broader range of intellectual tasks. The industry quickly shifted from building specialised tools to pursuing systems that could reason, write, code, design, analyse, and assist across multiple domains.
From automation tool to force multiplier
We have not reached artificial general intelligence yet, but in many areas we are already seeing systems that outperform average professionals on specific tasks. In software engineering, modern AI models can generate code, create user interfaces, analyse large codebases, identify defects, produce documentation, explain technical decisions, and accelerate development workflows that previously required teams of skilled specialists.
Human expertise still matters. In fact, it matters more than ever. The quality of the result depends heavily on the person defining the task, reviewing the output, setting constraints, and making final decisions. However, the nature of the work has changed. AI is no longer a toy or a research demonstration. It has become a genuine force multiplier for people and organisations that know how to use it properly.
This is why the current AI race is not simply another software trend. A technology that can amplify intellectual labour at scale has strategic importance. It affects productivity, competitiveness, cybersecurity, education, research, national security, and the structure of labour markets.
The genie is already out of the bottle
Only a few years ago, many observers expected advanced AI to become a tightly controlled technology. That would have been understandable. A system capable of dramatically increasing the productivity of developers, researchers, analysts, and engineers has obvious strategic value. It is not difficult to see why governments might treat such technology in the same category as cryptography, satellite systems, advanced semiconductors, or military-grade research.
Instead, much of the foundational research became publicly available. Open-source communities entered the field aggressively. Universities, startups, independent researchers, and global developer communities began building and improving their own models. Commercial providers continued to push the frontier, but open-source models advanced at a remarkable pace and narrowed the gap faster than many people expected.
Recent restrictions on access to some advanced models for non-US users may be the first visible sign that governments are beginning to view AI as a strategic asset rather than just another digital service. The problem is that this realisation may have arrived too late. The research exists. The expertise has spread. The open-source ecosystem is growing. Even if access to the most advanced proprietary models is restricted, alternative models will continue improving.
The security problem is bigger than the employment problem
Most public discussions about AI focus on jobs. That discussion is important, but it is not the only problem and may not even be the most dangerous one. From our perspective, the larger issue is security.
The world runs on an enormous amount of software. Much of it was written years ago. Some of it is poorly documented. Some systems depend on outdated libraries, abandoned components, and legacy infrastructure that is no longer actively maintained. Many vulnerabilities remain hidden simply because no one has had the time, money, or motivation to find them.
Advanced AI changes that equation. The same systems that help developers find and fix defects can also help malicious actors discover vulnerabilities, automate reconnaissance, analyse legacy systems, and develop attack strategies at scale. Criminal groups no longer need the same level of human expertise if they can use AI as a tireless technical assistant.
This risk extends far beyond software. AI can accelerate work in biology, chemistry, materials science, engineering, autonomous systems, and many other areas. The same technology that helps researchers and companies innovate can also help hostile actors explore harmful applications. For the first time, sophisticated intellectual capability is becoming cheap, scalable, and widely accessible.
The economics of AI adoption are distorted
There is another issue that receives less attention: pricing. Today, leading AI companies appear to be heavily subsidising access to advanced capabilities. Premium subscriptions often give users access to models and computational resources that may cost far more to provide than the subscription price suggests, especially when used intensively.
Some public estimates suggest that a power user on a $200 per month AI plan could theoretically consume many thousands of dollars’ worth of computational resources in a month. Whether the real number is $5,000, $8,000, or $14,000 is less important than the underlying point. Current pricing is designed to maximise adoption, build user habits, capture market share, and establish platform dominance. It does not necessarily reflect the long-term cost of delivering these services sustainably.
This has major consequences. Cheap AI accelerates adoption across businesses, freelancers, students, developers, and ordinary users. It also accelerates job displacement, market disruption, and security risks. When the cost of intellectual labour falls dramatically, the change affects everyone at once: legitimate companies, individual professionals, startups, criminals, and hostile state actors.
Why below-cost AI pricing matters
In many industries, aggressive below-cost pricing is treated as a competitive issue because it can distort markets. In AI, the problem is broader. Selling access to extremely powerful models at artificially low prices does not merely hurt competitors. It changes the speed at which society is forced to absorb a transformative technology.
If AI tools were priced closer to their real cost, adoption would still happen, but it would happen more gradually. Businesses would be forced to make more deliberate decisions. Some low-value use cases would not scale immediately. Certain jobs might have more time to evolve rather than disappear suddenly. Security teams would have more time to adapt. Governments and institutions would have a slightly wider window to prepare.
This would not stop malicious actors. It would not solve the open-source challenge. It would not remove the risks. But it could reduce the speed of uncontrolled diffusion and make the transition less chaotic.
Regulation alone will not be enough
Many policymakers naturally turn to regulation, but regulation alone is unlikely to solve the problem. Once a technology becomes globally distributed, digitally transferable, and technically reproducible, prohibition becomes extremely difficult to enforce. Restrictions often slow legitimate users more than determined actors operating outside the law.
The more realistic path is a combination of security investment, responsible deployment, economic scrutiny, infrastructure hardening, and international coordination where possible. Companies will need to audit legacy systems, improve software supply chain security, modernise critical infrastructure, and treat AI-enabled threats as a new baseline rather than a future possibility.
AI providers will also need to be more transparent about pricing, usage limits, safety controls, and the real economics of their services. If the current market is being built on massive subsidies, businesses should not assume that today’s prices will remain stable. The full cost of AI may eventually be passed on to customers, employers, workers, and society as a whole.
The real question
Artificial intelligence will bring enormous benefits. It will improve productivity, accelerate research, reduce the cost of software development, support education, and give smaller teams capabilities that once belonged only to large organisations. These opportunities are real and should not be dismissed.
At the same time, the risks are equally real. AI lowers the cost of creation, but it also lowers the cost of exploitation. It gives honest professionals better tools, but it also gives malicious actors better tools. It helps companies automate work, but it also accelerates the disruption of jobs and business models that were not prepared for this level of change.
The question is no longer whether AI will reshape industries, economies, and societies. It already is. The real question is whether businesses, governments, security teams, and institutions can adapt quickly enough to survive a transition that is moving faster than any previous technological revolution.
