By CoinAINews Staff |

OpenAI CEO Sam Altman says the two risks he worries about most as AI becomes more capable are surprisingly broad: humans losing control, and too much power concentrated in too few hands.

The comments, highlighted in a recent update from CoinMarketCap and attributed to a conversation with entrepreneur and podcast host David Senra, come as the AI industry moves toward increasingly capable and autonomous systems.

For Altman, the challenge is no longer simply whether AI can become more powerful. It is also about whether humans can remain in control of that power — and whether advanced AI becomes concentrated within a small group of companies, models or individuals.

Losing Control Is the First Major Risk

One of Altman's central concerns is what happens if AIsystems become capable enough that humans can no longer reliably control their behavior.

That does not necessarily mean a science-fiction scenario in which an AI suddenly takes over.

A more practical concern is whether increasingly capable systems could perform actions that developers did not intend, particularly as AI moves from generating answers to carrying out tasks with greater autonomy.

That question has become more relevant as AI companies develop systems capable of using tools, accessing digital environments and completing multi-step tasks with less direct human involvement.

The more autonomy an AI system receives, the more important monitoring, testing and containment become.

The Second Risk: Too Much Power in One Place

Altman's other concern is different.

It focuses not on what AI itself might do, but on who controls the most powerful AI systems.

If one company, one model or one individual gains disproportionate influence over advanced AI, the consequences could extend far beyond the technology sector.

Powerful AI could influence how people access information, how businesses operate and how important digital services are delivered.

That makes the distribution of AI capabilities an increasingly important part of the broader conversation.

The question is no longer simply who can build the best model.

It is also who gets to control it and who gets access to its capabilities.

Why Centralization Matters

Building frontier AI systems requires enormous computing resources, specialized chips, data, infrastructure and highly skilled researchers.

Those requirements naturally favor companies with substantial financial and technological resources.

There are advantages to that model. Large organizations can spend heavily on research, safety testing and infrastructure.

But concentration also creates potential risks.

If a small number of companies control the most capable systems, decisions made by those companies could have an outsized impact on businesses, consumers and governments.

That is why AI centralization has become a governance issue as much as a technology issue.

The AI Industry Is Already Debating This — and the Disagreements Are Becoming Sharper

Altman's concerns are not limited to OpenAI.

Other technology leaders have also entered the debate over how AI power should be distributed.

Meta CEO Mark Zuckerberg has argued that advanced AI should be broadly distributed rather than controlled by a small number of institutions. In an August essay, Zuckerberg described the distribution of AI capabilities as an important question for the future balance of power.

The disagreement highlights a fundamental divide within the industry.

Some argue that powerful AI needs to be developed carefully and controlled through strong safeguards.

Others emphasize wider access and competition, warning that excessive centralization could itself become a major risk.

Both sides are ultimately wrestling with the same question: how much control should any single organization have over increasingly powerful AI?

Safety vs. Innovation

The AI industry faces a difficult balancing act.

Companies want more capable systems because advanced AI could improve software development, scientific research, productivity and a wide range of other industries.

At the same time, greater capability can create greater risks if systems behave unexpectedly or gain access to sensitive tools and environments.

That creates pressure to improve safety measures alongside model capabilities.

OpenAI has previously described monitoring, alignment and security as important safeguards as AI systems become more capable.

The difficulty is that AI companies are also competing against one another.

A company that slows development significantly could risk falling behind competitors that continue moving faster.

That tension between speed and safety is likely to remain one of the defining issues of the AI industry.

AI Agents Make Control More Important

The question of control becomes even more important as AI moves beyond traditional chatbots.

AI agents can potentially use software, interact with websites, access information and complete tasks across multiple steps.

That makes them considerably more useful than systems that simply generate text.

It also increases the potential consequences of an error.

An AI that produces an incorrect paragraph may create a relatively limited problem.

An autonomous system with access to external tools could potentially take actions that affect real systems, accounts or data.

As a result, developers need stronger safeguards around permissions, monitoring and containment.

What Altman's Warning Means for AI

Altman's comments point toward a broader challenge facing the technology industry.

Building a more capable AI system is only one part of the problem.

Developers also need to consider how that system behaves, what it can access, who controls it and how its capabilities are distributed.

Those questions become more difficult as models become increasingly autonomous.

The industry may therefore be approaching a point where governance and control are nearly as important as raw model performance.

A model that is more powerful but difficult to control could create problems that cannot be solved simply by making it smarter.

The Bigger Picture

The AI debate is entering a new phase.

Early discussions focused heavily on whether AI could match or exceed human performance on particular tasks.

Today, the conversation increasingly involves questions about autonomy, safety, concentration of power and access.

Those issues will become even more important if AI systems begin handling larger portions of business operations, research and digital infrastructure.

For policymakers, developers and users, the challenge is finding a framework that encourages innovation without allowing powerful AI capabilities to become either uncontrollable or excessively concentrated.

The Bottom Line

Sam Altman's warning comes down to two fundamental risks: losing control over increasingly capable AI and allowing too much power to become concentrated in too few hands.

Neither problem has a simple solution.

AI development is advancing rapidly, while companies and governments are still working out how safety, competition and access should evolve alongside the technology.

The next stage of the AI race may therefore be about more than building the smartest model.

It may be about answering a much harder question: how do we make sure that increasingly powerful AI remains under meaningful human control — without letting its benefits become concentrated in the hands of too few?

This article is for informational purposes only and does not constitute financial, investment or technology-policy advice.