Sanders and Bannon Demand AI Put Humans First

older man in suit and glasses outdoors — Photo: Rich Koele / Shutterstock

When political opposites share a stage to demand that artificial intelligence remain a tool in human hands—not a lever of unaccountable power—you are seeing a governing coalition form around first principles, not party lines.

At a Glance

Washington’s “Pro-Human Assembly” brought together figures as far apart as Senator Bernie Sanders and Steve Bannon to argue for human-controlled AI with real guardrails.
The core program: transparency, testing, accountability, and limits on concentrated AI power—areas where bipartisan agreement is already measurable.
The event’s organizing network, centered on the Future of Life Institute, reflects years of advocacy for keeping AI development subordinate to human oversight and public consent.
The political alignment is durable because it aims at governance design—who sets rules, who enforces them, and how to avoid regulatory capture—not culture-war symbolism.

What Happened: A Cross-Ideological Case for Human Control

In Washington, D.C., the Pro-Human Assembly put an unusual coalition on the same program: U.S. Senator Bernie Sanders and Steve Bannon each took the podium to call for stronger guardrails on artificial intelligence and to frame AI as a governance problem—who controls it, who is accountable when it fails, and how to prevent a handful of firms from consolidating power through automation and scale. The event was explicitly billed as “pro-human”: AI should augment human capacity, remain subject to human direction, and be constrained where harms are clear. The organizers published the date, location, and roster, and scheduled the speakers in separate slots at the same conference, underscoring a shared objective despite stark ideological distance.

The rhetorical overlap was not about love of the same policies but recognition of the same risks: opacity, untested deployments, and market structures that invite dominance by a few vertically integrated AI providers. Reporting at the time captured both the improbable pairing and the substantive demand—tighter controls before systems outpace institutional capacity to govern them.

How This Coalition Coalesced: Narrow Consensus, Focused Aims

Cross-party AI governance has been building for years around a modest but durable center of gravity: require transparency (what a system is, how it behaves), mandate testing and evaluation before high-stakes use, keep a human decision-maker responsible in consequential contexts, and draw bright lines against clearly harmful applications. Analyses of congressional activity and bipartisan working groups have mapped these points of agreement repeatedly; the Senate’s early efforts to seed an AI framework reflected exactly this lane—structured briefings, joint listening sessions, and exploratory drafts aimed at transparency, testing, and oversight rather than grand redesigns of the tech economy.

That narrow lane is why the alliance holds. It does not depend on universal agreement about pauses, bans, or industrial policy. It depends on the logic that complex, adaptive systems deployed at scale must be observable, stress-tested, and auditable, and that where risk is systemic—think critical infrastructure, employment decisions, or lethal force—a human remains on the hook. Cataloging these principles into statute and standards is unglamorous work, but it is feasible work—and it is where the votes already are.

The Organizing Theory: Keep AI Subordinate to Human Institutions

The organizing network behind the Assembly, including the Future of Life Institute, has long argued a simple hierarchy: human values and public legitimacy first, technical advance second. Their campaigns, open letters, and resources frame AI as a fork in the road: one branch concentrates capability in a few unaccountable institutions; the other channels the technology to amplify human dignity, liberty, and community resilience. The point is not to arrest progress but to bind it to publicly acceptable ends, using governance, standards, and market rules to keep the direction pro-human rather than merely pro-efficiency.

That frame resonates because it translates abstract anxieties into institutional design problems. It asks concrete questions an appropriator or regulator can act on: What documentation must accompany a model? What evaluation regimes are mandatory before deployment in health care, finance, or education? Who bears liability when an automated system errs? And what structural safeguards—interoperability, data portability, separation of control over models and distribution channels—limit the chance that one company’s failure becomes everyone’s problem?

Where Agreement Ends: Scope, Speed, and Structure

Beyond the consensus core, disagreements sharpen. Some advocates endorse stronger brakes—conditional moratoria on frontier systems until specific safety criteria are met, or bans on classes of capability without robust controls. Others prefer a “safety-by-design plus disclosure” model and resist categorical prohibitions. The Future of Life Institute’s own campaigns have at times pressed for stringent measures, including calls to halt development toward superintelligence absent scientific and public consensus on controllability—language that places the burden of proof squarely on developers before pushing capability further.

There is also a structural debate about how to regulate without handing the pen to incumbents. Complex technology rules can be captured by the best-resourced firms, turning safety into a moat. Serious governance design tries to counter that: performance-based requirements rather than proprietary checklists, third-party evaluations rather than self-attestations alone, and multi-stakeholder standards bodies with clear public-interest representation. The risk of regulatory capture is not a trope—it is a well-studied failure mode in tech policy—and any pro-human regime that forgets it will entrench the very concentration it claims to resist.

SANDERS, BANNON CALL FOR EXECUTIVE ACTION ON AI RESTRICTIONS

Speaking at the “Pro-Human Assembly” in Washington today, progressive senator Bernie Sanders and rightwing strategist Steve Bannon have called for restrictions on artificial intelligence: “And this is not legislation.…

— Worldwide News Network (@WorldwideNNX) September 15, 2026

What Durable Pro-Human AI Looks Like in Practice

A workable pro-human framework does four things. First, it compels legibility: model and system cards that disclose provenance, training boundaries, evaluation results, and intended use; traceability for critical outputs; and audit trails that regulators and litigants can follow. Second, it demands pre-deployment testing calibrated to context—red-teaming for misuse, robustness checks against distribution shifts, and domain-specific safety cases in sectors like health or finance. Third, it fixes accountability in law: clear lines of responsibility among model developers, integrators, and deployers, and liability rules that do not evaporate in a maze of end-user license agreements. Fourth, it maintains market openness: interoperability requirements, data access governed by privacy-preserving mechanisms, and remedies against tying that would fuse compute, model access, and application distribution in ways that foreclose competition.

None of this requires unanimity about the outer edge of risk. It requires institutional muscle memory—agencies that can evaluate technical evidence, courts that can scrutinize causation in AI-mediated harms, and legislative clarity about thresholds that trigger stricter obligations. The bipartisan baseline already supports these pillars; what the Assembly adds is momentum and a narrative that places human primacy above technical bravura.

Why This Coalition Endures

Cross-ideological alignments often fade when the cameras leave. This one is more likely to persist because its payload is procedural power: who must disclose, who must test, who must answer when things go wrong. Those are evergreen questions, not tied to a single product cycle or election. The Assembly’s unusual roster was not the substance; it was the signal that governing the most general-purpose technology of our time cannot be left to those building it alone. As long as transparency, testing, human accountability, and anti-concentration remain the center of gravity, the pro-human coalition has a practical agenda and a map through Congress to pursue it.