The debate around artificial intelligence (AI) is no longer limited to the precision of models or their regulation. Questions of control, ownership and long-term autonomy now take center stage. In a recent publication on Not your keys, not your bots ”, crypto investor and author Balaji Srinivasan lays out a simple idea: he who controls the keys controls the machines.
In brief
- Srinivasan says AI still depends on humans to set its fundamental goals and direction.
- He doubts that AI can independently create its own long-term goals.
- True AI autonomy would require control of robots, drones and data centers.
- Blockchain keys could become the basis for future AI governance and control.
Srinivasan: AI still depends on humans to set its goals
Srinivasan asks the question simply: “The fundamental question is whether AI stays on a leash. » For now, he believes that humans remain in control. Even though AI can optimize instructions, analyze problems, and correct its own errors, it’s still a human who sets the goal. Decisions remain guided by markets, politics and developments in society.
However, he questions the duration of this balance. As AI systems advance and become better at verifying information and reasoning, they can perform many tasks better than humans. But it remains uncertain whether AI can define its own fundamental objectives. Srinivasan doubts it will stop depending on humans to give it short-term direction.
He relates this question to biology. Human motivations are rooted in evolutionary pressures like survival and reproduction. Unless AI systems can reproduce themselves without human intervention, he believes they will remain tied to the goals set by their creators or owners.
Governance could move to blockchain, says Srinivasan
The investor considers a more extreme scenario to test this hypothesis. Truly autonomous AI would, he said, require control of physical infrastructure without human supervision. This would include:
- Humanoid robots capable of acting autonomously.
- Drones operating without centralized human command.
- Data centers operating outside of traditional corporate or state structures.
- Energy systems and assembly lines managed entirely by machine coordination.
Such a development is not technically impossible “, he wrote, but geopolitical realities could determine the outcome. In China, for example, authorities would be more inclined to design systems that are strictly controlled rather than truly autonomous. In this model, robots and digital agents would be linked to human identities via cryptographic mechanisms.
Beyond China, Srinivasan sees blockchain cryptography as a possible governance layer for AI. Private property could increasingly rely on private keys. Bots and software agents would act as extensions of their owners, secured by cryptographic authentication. In this model, control of keys is equivalent to control of machines.
Independent, uncontrolled robots would be seen as security risks rather than major advancements. Compliant humans and machines would cooperate to prevent a system from replicating itself outside of an authorized framework.
Although speculative, Srinivasan's argument fits into broader debates about AI alignment, digital sovereignty, and infrastructure ownership. As governments develop new rules and companies develop ever-better models, his message to the crypto community is clear: securing keys means securing bots.
Maximize your Tremplin.io experience with our 'Read to Earn' program! For every article you read, earn points and access exclusive rewards. Sign up now and start earning benefits.
