Rent Seeking - not what I thought it was

I was reading Noahpinion Roundup #80 and I came across this...

Now, I've heard of "rent seeking" in the past, and my mind has always jumped to "Rent" as in pay the landlord some money. But this seemed like a weird use of the word... So I asked an AI what Rent Seeking means in this context, and my eyes were opened. I'm pasting the reply here, with slight formatting changes. This is not my writing, just a note.

Rent-Seeking in the Age of AI

What Is Rent-Seeking?

Rent-seeking is an economic concept originally formalized by economist Gordon Tullock (1967) and later named by Anne Krueger (1974). It refers to the practice of increasing one's share of existing wealth without creating any new wealth. In other words, it's about manipulating the economic or political environment to redirect income toward yourself, rather than earning it through productive activity.

The term "rent" here doesn't mean paying a landlord — it comes from the classical economic concept of economic rent: income earned in excess of what would be necessary to keep a resource in its current use.

Classic Examples of Rent-Seeking

• Lobbying for favorable regulations, subsidies, tariffs, or tax loopholes
• Regulatory capture — shaping rules to create barriers to entry that protect incumbents from competition
• Patent trolling — acquiring patents not to innovate, but to extract licensing fees or lawsuit settlements
• Frivolous or strategic litigation — using the legal system to extract settlements or exhaust competitors
• Occupational licensing that goes beyond public safety to restrict supply and inflate wages for insiders
• Zoning manipulation — influencing land-use rules to inflate property values

The key insight is that rent-seeking is legal. It's not theft or fraud — it operates within (and often shapes) the rules of the game. This is what makes it so insidious and hard to combat.

Why AI + Rent-Seeking Is a Uniquely Dangerous Combination

The extract you shared highlights a critical blind spot in AI discourse. Most conversations about AI risk fall into two buckets:

1. Productive uses — science, coding, medicine, automation
2. Criminal uses — bioterrorism, deepfakes, cyberattacks

But the third category — AI-powered rent-seeking — may be the most consequential, precisely because it's legal and therefore harder to regulate or even recognize. Here's how it could manifest:

1. AI-Powered Hyper-Lobbying

Imagine AI systems that can:

• Analyze every piece of pending legislation across all 50 states in real time
• Draft thousands of tailored comment letters for regulatory proceedings
• Identify the exact legislative language that would benefit a client, buried in a 1,000-page omnibus bill
• Optimize campaign contributions for maximum policy influence

This turns lobbying from an artisanal craft into an industrial-scale operation, massively amplifying the advantage of well-funded interests over the public.

2. Automated Legal Extraction

AI could supercharge:

• Mass patent trolling — using AI to generate or acquire vast patent portfolios and algorithmically identify infringement targets
• Strategic litigation at scale — filing enormous volumes of nuisance suits calibrated to be cheaper to settle than to fight
• Contract optimization — drafting consumer contracts, insurance policies, or terms of service that are maximally extractive while remaining technically legal

3. Regulatory Arbitrage and Gaming

AI systems could continuously scan regulatory environments to find:

• Loopholes in tax codes across jurisdictions
• Gaps between the letter and spirit of financial regulations
• Opportunities to structure transactions to extract subsidies, credits, or favorable treatment

This already happens — but AI could do it at a speed and scale that overwhelms regulators (who are typically under-resourced).

4. Algorithmic Price Manipulation

• AI pricing algorithms that achieve tacit collusion without explicit coordination (already under investigation in rental and airline markets)
• Dynamic pricing systems that extract maximum consumer surplus without technically violating antitrust law
• Sophisticated market microstructure exploitation in financial markets

5. Capture of the AI Regulatory Process Itself

Perhaps most meta: AI companies could use AI to shape the very regulations governing AI — writing model legislation, flooding comment periods, and outmaneuvering public interest groups.

Why This Is Especially Concerning

The fundamental problem is one of asymmetry: rent-seeking AI will be deployed by those with the most resources (corporations, wealthy individuals, sophisticated financial actors), while the costs are diffused across the general public in the form of higher prices, worse services, less competition, and a political system more captured by concentrated interests.

The Deeper Worry

Productive AI and criminal AI both have natural feedback loops — productivity gains are visible and celebrated; crimes are investigated and prosecuted. But rent-seeking operates in a gray zone where:

• The harm is diffuse (everyone pays a little more; no single victim)
• The activity is legal (so there's no enforcement mechanism)
• The beneficiaries are concentrated and motivated (and can reinvest rents into more rent-seeking)
• The complexity is overwhelming (the public can't follow what's happening)

AI doesn't just make rent-seeking more efficient — it could make it effectively invisible, buried in layers of algorithmic optimization that no human regulator can parse.

This is arguably one of the most important and under-discussed dimensions of AI risk: not that AI will destroy jobs or kill people, but that it will quietly redistribute wealth upward through perfectly legal means, at a pace and scale that democratic institutions are not equipped to counter.