AI & Technology — Innovation With Guardrails

The global AI market is projected to reach $1.3 trillion by 2032. Six companies — Google, Microsoft, Meta, Amazon, OpenAI, and Anthropic — control over 90% of frontier AI model development. The concentration is unprecedented: the entire trajectory of a technology that will reshape every sector of the economy is being determined by a handful of corporate boardrooms with zero democratic oversight.

AI is already making high-stakes decisions about American lives — and getting them wrong along predictable demographic lines. Amazon's hiring AI penalized résumés containing the word 'women's.' The COMPAS recidivism algorithm has a 45% false positive rate for Black defendants versus 23% for white defendants. HUD settled with Facebook over AI-driven housing ad discrimination. AI-powered tenant screening tools have denied housing based on flawed data. There is no federal law requiring audits, transparency, or accountability for any of these systems. See Issue #21 — Internet Privacy for the broader algorithmic accountability framework.

Deepfakes and AI-generated disinformation pose an existential threat to election integrity. AI-generated robocalls impersonated President Biden in the 2024 New Hampshire primary. Deepfake technology can produce indistinguishable video of any public figure saying anything. Only a handful of states have passed deepfake election laws; there is no federal prohibition. The FCC has ruled AI-generated voice calls illegal under existing robocall law, but enforcement infrastructure does not exist. See Issue #30 — Media & Press Freedom for the broader misinformation landscape.

AI is displacing workers at an accelerating pace. The World Economic Forum projected 85 million jobs would be displaced by automation and 97 million new roles created in the 2020–2025 period — but the displaced workers and the new roles are in different industries, different skill levels, and different geographies. There is no federal program for AI-specific workforce transition, retraining, or income support. See Issue #13 — Labor for the broader worker protection framework.

Federal and state governments are deploying AI in high-stakes decision-making with minimal oversight. AI tools assist in determining benefits eligibility, criminal sentencing recommendations, predictive policing, and immigration enforcement. An Idaho Medicaid algorithm cut home care hours for disabled residents by up to 42% before being struck down in court. No federal law requires transparency, human review, or due process protections for government AI decisions. See Issue #12 — Criminal Justice for AI in sentencing and policing.

For workers, the AI transition stops being a cliff and starts being a bridge. The millions of jobs facing automation pressure don't disappear overnight — the 180-day advanced notice requirement for AI-driven layoffs gives workers and communities time to prepare. The $20B AI Workforce Transition Fund creates portable retraining accounts, community college AI skills programs, and employer incentives for retraining existing workers rather than replacing them. The 97 million new roles the WEF projects become accessible to displaced workers, not just to new graduates.

For anyone subject to algorithmic decisions, the bias audit regime changes the game. Hiring AI that penalizes women's résumés gets caught before deployment, not after a Reuters investigation. Tenant screening algorithms that deny housing based on flawed eviction records face disparate impact standards. Lending algorithms that charge higher rates along racial lines are subject to mandatory third-party audits with published results. The private right of action means individuals can hold companies accountable — not just wait for an overwhelmed FTC.

For election integrity, AI-generated deepfakes of candidates become a federal crime within 90 days of an election. Mandatory disclosure labels on AI-generated media give voters the information they need to distinguish real from synthetic. Platform liability creates real enforcement incentives. The AI robocall impersonation that targeted New Hampshire voters in 2024 becomes prosecutable, not just FCC-rulemaking fodder.

For innovation, open-source AI survives. The NAIRR provides $500M+/year in public compute and research funding, ensuring that AI development is not a six-company monopoly. Public universities get guaranteed access to the compute resources that frontier AI requires. Open-weight model development is protected by statute — corporate lobbying cannot ban the open research that drives scientific progress. Competition survives consolidation.

For government accountability, every federal AI system is disclosed, audited, and subject to human review. The Idaho Medicaid algorithm that slashed disabled residents' home care hours by 42% — and was only caught through litigation — becomes impossible under these rules. Benefits determinations, sentencing recommendations, surveillance tools, and enforcement algorithms all require algorithmic impact assessments before deployment and annual public audits after. Due process applies to digital government, not just analog government.