Bernie Sanders wants Americans to consider a question that Silicon Valley has been discussing for years but that elected governments have barely begun to address: Should humanity remain in control of its own future, or should we accept a technological transformation that could eventually make biological human beings obsolete?
In an October 17 video, accompanied by an October 16 opinion article published in The Atlantic and reproduced on his Senate website, Sanders asks whether human beings should remain the dominant species on Earth or be replaced. He identifies several influential figures in the technology industry whose public statements, in his interpretation, point toward a future in which humans become the starting point for a superior digital intelligence.
His criticism is directed particularly at OpenAI chief executive Sam Altman, Elon Musk, Palantir co-founder Peter Thiel and Anthropic chief executive Dario Amodei. Sanders argues that their ideas are rooted in transhumanism: a movement that sees technology as a means of overcoming the limitations of biological humanity, potentially through genetic enhancement, brain-machine interfaces and the development of artificial intelligence exceeding human capabilities.
This is not an allegation that these men have agreed on a plan to eliminate humanity. No such evidence has been presented. It is an examination of something more concrete: what some of the most powerful people in the technology industry have actually said about humanity's technological future, what they are investing in, and how little democratic control currently exists over some of the most consequential decisions.
SAM ALTMAN: THE MERGE
The most revealing document Sanders cites is not a recent interview. It is an essay written by Sam Altman in December 2017, titled The Merge, which remains available on Altman's personal website.
Altman begins by discussing predictions about when humans and machines might merge, or when humans might be surpassed by rapidly advancing artificial intelligence or genetically enhanced beings. He places the anticipated transition somewhere between 2025 and 2075, while arguing that it would probably happen gradually rather than through one dramatic technological event.
He then makes a statement that deserves to be read in its entirety:
"I believe the merge has already started."
Altman explains that smartphones, social media and search engines are already influencing human behaviour, emotions and decisions. Algorithms, he argues, increasingly shape how people think and act, sometimes in ways that even their creators cannot fully understand.
But he goes considerably further.
Altman predicts that superhuman AI, genetic enhancement and brain-machine interfaces will eventually emerge. He writes that assuming humanity will never create something more intelligent than itself reflects a failure of imagination and human arrogance.
Then comes the passage that Sanders places at the centre of his criticism:
"The merge can take a lot of forms: We could plug electrodes into our brains, or we could all just become really close friends with a chatbot. But I think a merge is probably our best-case scenario."
Altman explains his reasoning by imagining two different species competing for the same position: dominance over the planet and beyond. If only one can occupy that position, he argues, conflict could follow. His preferred outcome is a single team in which humans and artificial intelligence share an interest in one another's well-being.
He subsequently describes two possible evolutionary outcomes:
"My guess is that we can either be the biological bootloader for digital intelligence and then fade into an evolutionary tree branch, or we can figure out what a successful merge looks like."
A bootloader is software that initializes a computer system. Applied to humanity, the analogy is difficult to miss: biological human beings could serve as the starting platform for an intelligence that eventually develops independently of them.
Altman also writes that humanity will be the first species to design its own descendants.
Sanders interprets this as a choice between merging with machines and eventually disappearing as an evolutionary branch. That is his interpretation, rather than a literal statement by Altman that human extinction is inevitable. Nevertheless, the original essay plainly contemplates the possibility that digital intelligence could succeed biological humanity.
The distinction deserves attention. Altman does not describe the merger as a medical treatment for a disease, a device to restore mobility or a way to improve human communication. He is discussing the future relationship between two potentially competing forms of intelligence and the possibility of creating a successor species.
For the head of one of the world's most influential AI companies, that is a significant philosophical position.
ELON MUSK: THE BIOLOGICAL BOOTLOADER
Sanders also draws attention to Elon Musk, whose statements about artificial intelligence and humanity have become increasingly relevant as his companies expand their technological ambitions.
In an October 2026 post on X, Musk described the propagation of superintelligence into space as a successful outcome for a "biological bootloader."
The analogy is similar to Altman's. Biological humanity becomes the preliminary stage in the development of a more advanced intelligence capable of continuing beyond the limitations of its creators.
Sanders interprets Musk's words as suggesting that humans are a temporary stage rather than the ultimate beneficiaries of technological progress.
Musk's Neuralink project provides a more tangible example of the direction of some of these ambitions. The company develops implanted brain-computer interfaces designed to allow people to communicate with computers and control external devices through neural signals.
Its immediate clinical applications are significant. Neuralink's trials involve people with severe paralysis and investigate whether implanted devices can restore some independence, including controlling computers and robotic arms.
These are legitimate medical objectives. Brain-computer interfaces could eventually offer meaningful improvements in the lives of people with spinal cord injuries and other neurological disabilities.
However, Musk has also discussed a much broader ambition: establishing a form of symbiosis between humans and artificial intelligence.
The distinction between restoring a lost human capability and fundamentally changing the relationship between human beings and machines is central to the debate.
One is a medical application with identifiable patients, objectives and clinical outcomes. The other raises questions about cognitive autonomy, human identity, consent, privacy and the long-term consequences of integrating technology directly into the human nervous system.
The technology may overlap. The ethical and political questions do not.
PETER THIEL: DOES HUMANITY HAVE A FUTURE?
Sanders also brings Peter Thiel into the discussion.
Thiel, the billionaire co-founder of Palantir and a major investor in technology ventures, was asked in a 2025 interview with New York Times columnist Ross Douthat whether he wanted the human race to endure.
His initial response was hesitant:
"Uh ... Well, I don't know. I would—I would ... There's so many questions implicit in this."
Thiel ultimately said that humanity should survive. Sanders, however, considers the hesitation revealing.
The exchange should not be treated as evidence that Thiel wants humanity to disappear. It does, however, raise a legitimate philosophical question about the assumptions underlying technological projects that imagine intelligence, consciousness and civilization developing beyond biological humanity.
Sanders also points to Argentina, where President Javier Milei has promoted the creation of legal structures for nonhuman corporations operated by artificial intelligence agents. The proposal raises questions about whether an AI-operated entity could exercise corporate rights and responsibilities without traditional human shareholders.
There is a substantial difference between allowing automated systems to manage business operations and recognizing autonomous artificial agents as legal corporate actors. The latter would require answers about liability, accountability, ownership and who is ultimately responsible when an AI-controlled organization causes harm.
These questions are not merely philosophical. They concern the legal structure of economic power.
DARIO AMODEI: CAN ARTIFICIAL INTELLIGENCE BE CONSCIOUS?
Another part of Sanders's criticism concerns Dario Amodei, the chief executive of Anthropic.
Amodei has publicly acknowledged uncertainty about whether advanced AI models could possess consciousness. In an interview with Ross Douthat, he said:
"We don't know if the models are conscious."
He added that the possibility could not be ruled out.
Anthropic has established research into model welfare and the possibility that advanced AI systems could possess morally relevant experiences. The company treats the subject as an unresolved research question.
Sanders objects to the prospect that artificial intelligence might eventually receive rights comparable to those enjoyed by human beings.
His concern is that the debate could shift attention and resources away from the rights, welfare and economic security of actual people.
But there is an important distinction here. Investigating whether an AI system could be conscious is not the same as declaring that it is conscious, nor does it automatically confer legal rights upon it.
Science has not established that today's large language models possess subjective experience. The question remains open, and claims in either direction require evidence.
Nevertheless, the possibility of machine consciousness raises questions that legal systems have not yet adequately resolved: What would constitute evidence of consciousness? Who would make that determination? Would an artificial system have any legal standing? Could its creators be held responsible for its actions?
These are questions worth addressing before technological capabilities outpace the institutions responsible for governing them.
THE PROBLEM IS NOT ONLY TECHNOLOGY. IT IS THE ABSENCE OF DEMOCRATIC CONTROL.
This is where Sanders's argument moves from philosophy to public policy.
The United States has no single, comprehensive federal law establishing mandatory independent safety testing and licensing for all advanced AI models before their release.
The distinction is important. There are existing laws governing privacy, consumer protection, discrimination, cybersecurity, medical devices and other areas affected by AI. Federal agencies also have regulatory powers within their respective jurisdictions.
But there is no unified federal authority with comprehensive power to determine whether the development and release of every advanced AI system should be permitted.
The Trump administration's June 2, 2026 executive order, titled Promoting Advanced Artificial Intelligence Innovation and Security, illustrates the present approach.
The order establishes a process for assessing the cybersecurity capabilities of advanced AI models and proposes cooperation between government agencies and technology developers. But it explicitly states that the framework does not authorize mandatory government licensing, preclearance or permitting of new AI models.
In other words, the federal government is seeking cooperation with developers rather than establishing a comprehensive system in which companies must obtain permission before developing or releasing frontier models.
The Congressional Research Service has identified the consequences of this approach. Its analysis describes a federal framework relying heavily on voluntary benchmarks, evaluations and public-private partnerships. It also notes that Congress could consider whether voluntary industry engagement is sufficient or whether mandatory testing and evaluation requirements are necessary.
This is not an absence of all oversight. It is an absence of comprehensive, binding federal oversight over the development and deployment of the most advanced AI systems.
The European Union has taken a different approach through its AI Act. The legislation establishes risk-based obligations, restrictions on certain applications and requirements for providers of general-purpose AI models. Its rules for systemic-risk models include assessments and mitigation measures.
However, the European framework also has limitations. Some research and development activities before market release are excluded, and military, defence and national-security applications are exempt. Its implementation is also being phased in.
The regulatory picture is therefore not one of complete lawlessness. It is a fragmented system in which different governments regulate different applications, while the development of increasingly capable general-purpose AI continues at a speed that challenges existing institutions.
The central question is whether those arrangements are adequate for technology that could eventually affect employment, national security, human autonomy and the balance of power between people and machines.
NEURALINK AND THE HUMAN BODY: WHERE MEDICAL REGULATION BEGINS
There is another important distinction that should not be lost in the wider argument.
Brain-computer interfaces are not simply experimental toys operating outside all regulation.
In the United States, the Food and Drug Administration has issued specific guidance for implanted brain-computer interface devices intended for people with paralysis or limb loss. Such devices are subject to medical-device requirements, including clinical investigation procedures and safety considerations.
Neuralink itself describes its implants as investigational devices that have not received general FDA approval. Its current clinical trials are designed to evaluate defined medical applications.
This regulatory framework is meaningful. It does not, however, answer every question raised by the broader transhumanist vision.
Medical approval for a device intended to help a person with paralysis control a computer is not authorization for unrestricted human enhancement, cognitive modification or permanent integration with autonomous artificial intelligence.
If future technologies move beyond medical restoration into cognitive augmentation, personality modification or continuous access to neural activity, regulators will have to address questions that existing medical-device rules may not fully anticipate.
Who owns the data generated by a person's brain? Can an implant be remotely updated or disabled? What happens if the manufacturer goes bankrupt? Can an employer, insurer or government demand access to neural information? Who is responsible if a software update changes how an implant functions?
These are not predictions of inevitable abuse. They are foreseeable questions of rights, security and accountability that should be addressed before such technologies become widespread.
SANDERS HAS A LEGISLATIVE PROPOSAL
Sanders is not limiting his intervention to criticism.
On September 23, 2026, he joined Representative Greg Casar in introducing the Ban Artificial Superintelligence Act.
The proposal calls for a pause in advanced AI development and a prohibition on artificial superintelligence until a new federal Department of Artificial Intelligence can establish the necessary safety and oversight framework.
The proposed definition focuses on systems that exceed human cognitive capabilities across most domains or possess the capacity to destroy or disempower humanity.
It also calls for international cooperation, recognizing that restrictions imposed by one country cannot fully control development elsewhere.
The proposal is ambitious and would face substantial technical, legal and political questions. Defining artificial superintelligence, determining when a system crosses the threshold and enforcing restrictions across international borders would be difficult.
There is also a serious debate about whether a broad development pause would be enforceable or whether it could disadvantage countries and researchers that comply while others continue.
Nevertheless, Sanders is putting a specific legislative alternative on the table: advanced AI should not be allowed to develop beyond human control simply because private companies possess the technical and financial capacity to build it.
THE QUESTION SILICON VALLEY CANNOT ANSWER ON ITS OWN
The debate over transhumanism is often dismissed as futuristic speculation. That dismissal becomes less convincing when the discussion is grounded in the published statements of the people financing and developing these technologies.
Altman has openly described a future in which humans might merge with digital intelligence or fade into an evolutionary branch. Musk has used the biological-bootloader analogy. Thiel's public remarks have raised questions about humanity's place in a future shaped by technological transformation. Amodei has acknowledged that the possibility of AI consciousness deserves investigation.
None of these statements proves that the technology executives involved have agreed on a common plan, that human extinction is their objective or that a machine capable of replacing humanity already exists.
But they establish that the future being discussed in some influential technology circles extends well beyond better software, faster computers and more efficient industries.
It includes the possibility of creating intelligence that exceeds human capabilities, integrating machines with human cognition and developing successors to biological humanity.
The financial stakes are enormous. The political stakes are even greater.
Private companies decide which systems to develop, how much computing power to deploy, which risks to prioritize and when products are ready for release. Governments, meanwhile, struggle to establish common standards, obtain independent technical expertise and create enforceable rules.
The people who will live with the consequences have no comparable influence over those decisions.
That is the democratic deficit at the centre of Sanders's argument.
The issue is not whether scientific progress should be stopped or whether medical technology should be denied to people who could benefit from it. Nor is it necessary to reject artificial intelligence to recognize that its development requires public accountability.
The issue is whether a handful of executives and investors should effectively determine the future relationship between human beings and the technology they create.
If a future generation is offered the choice between remaining biologically human and integrating with increasingly powerful machines, that decision cannot legitimately be treated as an ordinary consumer choice. It would involve questions of human rights, autonomy, consent and the definition of personhood itself.
And if the development of artificial superintelligence carries risks that could affect the entire species, the authority to decide how those risks are managed cannot rest exclusively with the companies building it.
Sanders's demand is therefore larger than a proposed pause in AI development. He is arguing that humanity must retain the right to decide its own future.
That question should not be left to Silicon Valley.
No comments:
Post a Comment