Tag: big tech in control

8.  Role Reversal—Humans As Fact-Checkers

Illustration by James Steinberg for TIME

Experts heavily emphasize the need for “human-in-the-loop” systems, ensuring that AI acts as a tool for assistance rather than an autonomous decision-maker, particularly in high-stakes fields like medicine and law. But as AI becomes more sophisticated, the number of “humans in the loop” narrows significantly. The future effect of AI is not preordained but will depend on the regulatory and ethical guardrails society puts in place, according to Brookings and Pew Research Center. 

What happens when AI becomes advanced enough to outpace human oversight? This gap, known as the alignment problem, occurs when AI systems become so complex and make decisions based on patterns so vast that humans can no longer verify every step of their reasoning. Researchers pursued one aspect of this question in 2021, asking “by 2030, will most of the AI systems being used by organizations of all sorts employ ethical principles focused primarily on the public good?”

In response, 68% chose the option declaring that ethical principles focused primarily on the public good will not be employed in most AI systems by 2030; 32% chose the option positing that ethical principles focused primarily on the public good will be employed in most AI systems by 2030.

  • Worries: The main developers and deployers of AI are focused on profit-seeking and social control, and there is no consensus about what ethical AI would look like. Even as global attention turns to the purpose and impact of artificial intelligence (AI), many experts worry that ethical behaviors and outcomes are hard to define, implement and enforce. They point out that the AI ecosystem is dominated by competing businesses seeking to maximize profits and by governments seeking to surveil and control their populations. It is difficult to define “ethical” AI: Context matters.

There are cultural differences, and the nature and power of the actors in any given scenario are crucial. Norms and standards are currently under discussion, but global consensus may not be likely. In addition, formal ethics training and emphasis is not embedded in the human systems creating AI.

  • Control of AI is concentrated in the hands of powerful companies and governments driven by motives other than ethical concerns: Over the next decade, AI development will continue to be aimed at finding ever-more sophisticated ways to exert influence over people’s emotions and beliefs in order to convince them to buy goods, services and ideas.
  • The AI genie is already out of the bottle, and abuses are already occurring. Some are not very visible and hard to remedy: AI applications are already at work in “black box” systems that are opaque at best and, at worst, impossible to dissect. How can ethical standards be applied under these conditions? While history has shown that when abuses arise as new tools are introduced, societies always adjust and work to find remedies; this time it’s different. AI is a major threat.
  • Global competition, especially between China and the U.S., will matter more to the development of AI than any ethical issues: There is an arms race between the two tech superpowers that overshadows concerns about ethics. Plus, the two countries define ethics in different ways. The acquisition of techno-power is the real impetus for advancing AI systems. Ethics takes a back seat. [1]

Hopes cited by study participants focused on inevitable advances in the technology that would include ethics “that will come as different fields show the way.” Currently, AI ethical regulations are shifting from voluntary guidelines to binding laws, focused on transparency, bias mitigation, and safety. The EU AI Act (2024) is the most comprehensive, using a risk-based approach, but in the U.S., regulation is currently fragmented, relying on Executive Orders, the NIST AI Risk Management Framework, and sector-specific enforcement against unfair/deceptive practices. The EU AI Act classifies AI by risk level (unacceptable to low), banning certain practices and imposing strict compliance for high-risk systems.

A U.S. Executive Order 14110 passed under President Biden in 2023 mandates safety tests for powerful models and directs federal agencies to set AI standards. In 2026, the Trump Administration is moving away from the previous focus on AI “equity” and “algorithmic fairness” toward a “minimally burdensome” national policy framework designed to accelerate innovation and American AI dominance.

Key actions and standards passed/ordered to date include Core AI Policy Framework & Orders (2025–2026), a four-page blueprint urging Congress to pass a unified federal AI standard, focusing on seven pillars, including child protection, intellectual property, and preemption of state AI laws under Executive Order 14365. A National Policy Framework for AI (Dec 11, 2025) attempts to centralize AI regulation at the federal level and reduce the “patchwork” of state regulations.

It includes:

  • AI Litigation Task Force: Established within the Department of Justice to challenge state AI laws that the administration deems to hinder innovation.
  • Conditional Funding: The order directs the Department of Commerce to potentially withhold Broadband Equity Access and Deployment (BEAD) funding from states that implement “onerous” AI regulations.
  • “Preventing Woke AI” Executive Order (July 23, 2025): This order prevents the federal government from procuring or using AI systems that contain ideological biases (such as, for example, producing inaccurate historical depictions). [Contrary to this idea, because AI is trained on historical human data, it can inadvertently perpetuate or amplify racial, gender, and socioeconomic biases. It might learn to favor male candidates because more men were historically hired in those roles. Language models learn meaning by analyzing how words are used together. If society associates certain professions (like “nurse”) with women and others (like “doctor”) with men, the AI simply mirrors that pattern.]
  • Removing Barriers to American Leadership in AI (Jan 23, 2025): An early order focusing on enhancing global AI dominance by streamlining regulations and promoting human flourishing.

Key Aspects of the Trump Administration’s AI Standards Preemption of State Laws:

  • The administration is pushing to have federal standards supersede state laws, such as those passed in California or Colorado, which it deems “innovation-limiting.”
  • “Truthful” AI Standards: New policies emphasize that AI should be free from DEI (Diversity, Equity, and Inclusion) requirements that the administration argues can cause AI to produce delusional or biased outputs.
  • Sector-Specific Regulation: Rather than an all-encompassing AI regulatory body, the framework favors a decentralized model where existing federal agencies handle regulation within their sectors.
  • Infrastructure Support: The orders aim to accelerate federal permitting for data centers to ensure the power needed for AI development.
  • General Data Protection Regulation (GDPR): Affects AI via strict privacy, automated decision-making, and profiling rules.
  • State-Level U.S. Laws: States like California (CPRA) are passing laws restricting automated decision-making and enhancing data privacy.
  • National Institute of Standards and Technology AI Risk Management Framework: A leading voluntary framework in the U.S. providing guidelines to map, measure, and manage AI risks

Resistance to the Trump administration’s 2025–2026 Artificial Intelligence (AI) standards—which prioritize a light-touch regulatory approach, rapid infrastructure development, and federal preemption of local rules—has emerged from a coalition of state lawmakers, bipartisan federal legislators, and tech companies concerned about safety and corporate accountability. Vorys Law Firm recently released the following commentary on this point.

States have begun responding more directly to the executive order, both through continued legislative activity and public statements from state officials. Notably, several state-level AI statutes have taken effect or are scheduled to take effect in 2026.  This includes California’s AI Transparency Act and Texas’s Responsible Artificial Intelligence Governance Act, both of which incorporate privacy focused provisions and impose disclosure and governance requirements related to transparency, automated decision making and how personal data can be used to train or operate AI systems.  Colorado’s comprehensive AI legislation is also scheduled to take effect on June 30, 2026, further expanding state-level regulation of high-risk AI systems.

In parallel, there is no indication that state legislative activity is slowing in response to the federal initiative.  Multiple states continue to advance AI-related bills toward enactment, including Washington, where several AI-related measures have progressed; Florida, where proposed legislation seeks to establish AI-related consumer rights; Virginia, where legislation addressing the use of AI in mental health applications has advanced; and Utah, where amendments addressing AI transparency requirements have moved through the legislative process. 

These developments reflect sustained momentum at the state level despite the Trump Administration’s emphasis on federal coordination.  By asserting federal primacy, privacy protections embedded in state frameworks could be narrowed or displaced.  While this may reduce compliance fragmentation, a federal overhaul may also eliminate specific data-handling standards that currently guide AI governance practices. [2]

The administration has targeted state-level AI safety bills, such as the Colorado AI Act, which the White House labeled “onerous” for aiming to prohibit algorithmic discrimination. In Republican-led states like Louisiana and Missouri, AI bills have stalled due to pressure from the White House, causing tension within the party over state rights vs. federal oversight.

On March 20, 2026, Democratic lawmakers introduced the GUARDRAILS Act, legislation to specifically block President Trump’s December 2025 executive order from taking effect, aiming to preserve state authority to adopt AI safeguards. Also, the AI Data Center Moratorium Act, introduced by Sens. Bernie Sanders and Rep. Alexandria Ocasio-Cortez, seeks to pause the construction and expansion of AI data centers until national, robust safety standards are in place. Many lawmakers are opposed to the federal government’s attempt to “steamroll” state initiatives, viewing the White House’s agenda as an attempt to grant “AI amnesty” to large tech companies.

The administration’s aggressive, pro-growth approach has led to high-profile conflicts with AI firms refusing to adhere to unregulated, military-driven, or “ideologically” driven standards. The Trump administration ordered federal agencies to cease using AI from Anthropic, a major U.S. firm, after it refused to allow its technology to be used for mass surveillance or autonomous, non-human-controlled weapons. Following the dispute, the Department of Defense (Department of War) designated Anthropic as a supply chain risk, a move criticized by tech lobbyists and lawmakers as a bullying tactic. The mind boggles at the underlying intent of government using AI for autonomous, non-human-controlled weapons, not to mention mass surveillance.

Already, in this mild instance alone, we see the potential for harm from AI under the control of a bad actor.

Critics argue the administration’s focus on “anti-woke” AI—which requires technology to be neutral and prevents ideological bias in outputs—is actually a push for ideologically skewed models that ignore real-world discrimination risks. Legal experts warn that the administration’s efforts to stop state regulation are deeply misguided and likely to face legal battles, as the Supreme Court has recently limited federal agency deference. Reports suggest that the administration’s framework neglects critical risks such as data privacy, algorithmic bias, and environmental impact in favor of accelerating development.

Observers need not look far to know who is behind the Trump Administration’s aggressive non-regulatory push. A question to Google AI listed the following culprits:

  • David Sacks: White House AI and Crypto Czar and co-chair of the President’s Council of Advisors on Science and Technology (PCAST). He has been pivotal in designing the 2025/2026 Action Plan, focusing on winning the AI race through deregulation.
  • Michael Kratsios: Former Chief Technology Officer in the first Trump administration and later co-chair of PCAST. He is a primary architect of the administration’s “permissionless innovation” approach, which seeks to reduce regulatory burdens on AI developers.
  • Elon Musk & Sriram Krishnan: Served as informal advisors, influencing the policy agenda to favor deregulation and align with the “tech right” movement in Silicon Valley.
  • Howard Lutnick: Commerce Secretary, identified as a key architect of economic policies related to technology, including tariff threats to influence industry standards.
  • Tech Leaders and Industry Influence (2025–2026 PCAST) The 2025–2026 Trump administration has explicitly brought industry leaders into the policy-making process through the revived PCAST, including
    • Mark Zuckerberg (Meta): Joined the tech council to help shape AI standards.
    • Jensen Huang (NVIDIA): A key advisor who has worked to align federal policy with the need for high-performance chip availability.
    • Larry Ellison (Oracle): A prominent supporter of the administration who has advised on AI infrastructure, with Oracle actively testing AI tools for federal use.
    • Sergey Brin (Google Co-founder): Appointed to the PCAST.
    • Marc Andreessen (Andreessen Horowitz): A venture capitalist whose firm has played a significant role in influencing Republican AI policy.
  • Key Interests and Ideology “Tech Right” & Deregulation: The policy is heavily influenced by a “tech-right” faction of venture capitalists and entrepreneurs who oppose “woke AI” and argue against diversity, equity, and inclusion (DEI) constraints in AI programming.
  • Anti-State Patchwork: A primary interest is preventing states, especially California and Colorado, from enforcing strict AI regulations. The administration aims to create a single national policy that preempts state-level rules.
  • National Security and Economic Dominance: The central goal is to ensure American companies, rather than Chinese competitors, lead in building AI infrastructure. This includes promoting “full-stack” American AI technology exports.
  • Infrastructure Development: Key stakeholders are pushing for accelerated permitting for data centers and increased power supply to feed AI systems.

Experts estimate the odds of humans permanently maintaining comprehensive control over super-intelligent AI at roughly 50% or less, with many viewing it as an existential challenge rather than a solved problem. The difficulty scales directly with AI advancement: as systems approach human-level capabilities, the complexity of their internal reasoning outpaces human cognitive capacity to continuously monitor.

[1] Pew Research Center, June 16, 2021. https://www.pewresearch.org/internet/wp-content/uploads/sites/9 /2021/06/PI_2021.06.16_Ethical-AI-Design_FINAL.pdf   “Pew Research and Elon’s Imagining the Internet Center built a database of experts to canvass from a wide range of fields, choosing to invite people from several sectors, including professionals and policy people based in government bodies, nonprofits and foundations, technology businesses, think tanks and in networks of interested academics and technology innovators. The predictions reported here came in response to a set of questions in an online canvassing conducted between June 30 and July 27, 2020. In all, 602 technology innovators and developers, business and policy leaders, researchers and activists responded to at least one of the questions covered in this report.”

[2] Listed as “Related Professionals”: Ingram, Christopher; LaRocco, Christopher; Needle, Celina. “Battle for AI Governance: White House’s Plans to Centralize AI Regulation and States’ Continuous Opposition,” NORYS, Apr 28, 2026. https://www.vorys.com/publication-battle-for-ai-governance-white-houses-plan-to-centralize-ai-regulation-and-states-continuous-opposition

Excerpted from Artificial Intelligence: A Pandora’s Box