Side-by-side analysis of what each approach would mean for AI regulation, deepfakes, algorithmic bias, worker protection, open source, and whether machines should make life-or-death decisions.
We're a policy platform with 50 researched positions on every major issue. This page compares AI policy approaches — but there's much more to explore.
Artificial intelligence is the most transformative technology since the internet — and it's advancing faster than any government can keep up. AI systems are already making decisions about who gets hired, who gets loans, who gets paroled, what medical treatment you receive, and what information you see online. Deepfake technology can fabricate convincing video of anyone saying anything. And the global AI arms race — with the US and China as the primary competitors — is reshaping military capabilities and economic power.
Yet the United States has no comprehensive AI law. No federal regulation governs algorithmic bias in hiring. No federal law requires disclosure when AI makes consequential decisions about your life. No federal statute addresses deepfakes comprehensively. The gap between AI's power and AI's governance is the defining policy challenge of this decade — and both parties are failing to address it.
Democrats favor cautious regulation through executive action. Republicans favor minimal regulation to encourage innovation. The Common Good Party proposes a risk-based federal framework: light regulation for low-risk AI, strict oversight for high-risk applications, and clear prohibitions on the most dangerous uses — while investing in American AI leadership and protecting workers through the transition.
How the three approaches stack up on AI and technology policy.
| Issue | Democrats | Republicans | Common Good |
|---|---|---|---|
| Federal regulation | Executive orders, voluntary frameworks | Minimal — innovation first | Risk-based federal law: light/strict/prohibited tiers |
| Transparency | Support disclosure requirements | Voluntary industry standards | Mandatory disclosure for consequential AI decisions |
| Bias auditing | Support, some proposed legislation | Oppose mandates, self-regulation | Mandatory independent audits for high-risk AI |
| Deepfakes | Support regulation, some state laws | Free speech concerns, targeted approach | Federal law: labeling, criminalize non-consensual, election protection |
| Worker protection | Retraining programs, some regulation | Market adjustment, oppose mandates | Transition fund, 12-month notice, portable benefits |
| Open source | Generally support, some safety concerns | Support — reduce big tech power | Support with safety guardrails for frontier models |
| Antitrust | Enforce existing law, some new proposals | Concerned about big tech censorship | Prevent AI monopolies, data portability, interoperability |
| Safety research | Fund AI safety, NIST standards | Industry-led, minimal government role | $5B federal AI safety institute, pre-deployment testing |
| Military AI | Human oversight required for lethal decisions | Maintain US advantage, fewer restrictions | Human control over lethal force, international framework |
| International cooperation | Multilateral frameworks, G7 coordination | US leadership, bilateral agreements | Democratic AI alliance, global safety standards |
Sources: NIST, White House OSTP, Congressional AI Caucus, party platform documents. See the compact comparison view for a quick summary.
Democrats have addressed AI primarily through executive action rather than legislation. The Biden AI Executive Order established safety testing requirements, watermarking standards, and reporting obligations for frontier AI models. Democrats support algorithmic transparency, bias auditing, and worker protection. The party has funded AI safety research through NIST and NSF, supported the CHIPS Act to boost domestic AI chip production, and pushed for international AI governance through G7 and bilateral frameworks.
Democrats correctly recognize that AI governance cannot wait for the technology to mature — by then, harms will be entrenched. The executive order established important precedents on safety testing and transparency. Investment in AI safety research is critical. And the emphasis on international coordination reflects the global nature of AI development — unilateral regulation alone is insufficient.
Executive orders are not legislation. They can be rescinded by the next president, creating regulatory uncertainty that undermines both safety and innovation. Democrats have not passed comprehensive AI legislation despite controlling Congress and the presidency. The party's approach relies too heavily on voluntary industry commitments — which companies honor when convenient and ignore when profitable. And the emphasis on regulating AI's risks has not been matched by a positive vision for how AI can benefit workers, healthcare, education, and public services.
For more on AI governance, see the full AI explainer.
Republicans generally oppose comprehensive AI regulation, arguing it would stifle innovation and cede competitive advantage to China. The party favors industry self-regulation, voluntary safety commitments, and minimal government intervention. Some Republicans support targeted measures — particularly on deepfakes in elections and Chinese AI threats — but oppose broad regulatory frameworks. The party emphasizes maintaining US AI dominance as a national security priority and has sought to rescind or weaken the Biden AI Executive Order.
The innovation concern is not trivial. Poorly designed regulation can indeed slow development, create compliance burdens that favor large companies over startups, and push AI development to less regulated jurisdictions. Maintaining US AI leadership over China is a legitimate national security priority. And open-source AI models — which some Democrats want to restrict — provide important benefits for competition, transparency, and democratized access to technology.
Industry self-regulation has never worked for powerful technologies with significant externalities. The social media era demonstrated exactly what happens when technology companies self-regulate — they optimize for engagement over safety, ignore documented harms, and resist accountability until forced. AI is far more powerful than social media. Allowing AI systems to make consequential decisions about people's lives without transparency, auditing, or accountability is not pro-innovation — it's pro-discrimination, pro-opacity, and anti-consumer.
The "regulation helps China" argument is also backwards. China has more AI regulation than the US — including deepfake laws, algorithmic recommendation rules, and generative AI regulations. The regulatory vacuum in the US doesn't give America an advantage; it creates an environment where harms go unaddressed and public backlash eventually produces worse regulation than thoughtful, proactive governance would.
For more on the regulation debate, see our AI explainer.
The Common Good Party proposes federal AI legislation — not executive orders — built on a risk-based framework. Low-risk AI (chatbots, recommendation engines, creative tools): minimal regulation, transparency labeling. High-risk AI (hiring, lending, healthcare, criminal justice, insurance): mandatory transparency, independent bias audits, human oversight, right to contest decisions. Prohibited AI: social scoring systems, real-time mass surveillance, fully autonomous lethal weapons without human control. Additionally: a $5 billion federal AI safety institute, a workforce transition fund for displaced workers, a federal deepfake law, antitrust enforcement to prevent AI monopolies, data portability requirements, and an international democratic AI alliance for coordinated governance.
Unlike the Democratic approach, we propose legislation rather than executive orders — because governance that can be rescinded overnight is not governance. Unlike the Republican approach, we recognize that unregulated AI is not innovative AI — it's dangerous AI that will eventually produce a backlash far more restrictive than proactive regulation. Our risk-based framework is designed to be pro-innovation for the vast majority of AI applications while providing real accountability for the applications that affect people's rights and safety.
The EU's AI Act, while imperfect, demonstrates that comprehensive AI regulation is feasible without destroying innovation. Japan, South Korea, Canada, and the UK are all developing AI governance frameworks. The US is the outlier in having no comprehensive approach. Meanwhile, the costs of ungoverned AI are already visible: algorithmic discrimination in hiring and lending, deepfake-driven fraud and disinformation, and the beginning of AI-driven job displacement without adequate worker protection. Acting now — while the technology is still developing — is far better than waiting for a crisis to force reactive, poorly designed regulation.
AI is already making decisions about your life. Here's what the CGP plan means for real people.
Want to explore how the full Common Good platform addresses technology and workers? See our policies on cybersecurity, labor, and internet privacy.
Explore the Full PlatformCommon questions about AI and technology policy.
Have a question not answered here? Read the full AI explainer or visit our site-wide FAQ.
Dive deeper into AI and technology policy.
No federal AI law. No deepfake protections. No worker transition plan. Read the full plan and see which approach governs AI before AI governs us.
Paid for by The Common Good Party (thecommongoodparty.com) and not authorized by any candidate or candidate's committee.