Dissident Politics: Why drilling down to get inconvenient truth is usually hard to impossible

People with wealth and power have a great information asymmetry advantage over the public and the public interest. The asymmetry poisons AI (artificial intelligence), making getting to real truth a very serious problem. This is a follow-on post to the one I posted earlier this morning. This is a lot shorter and has just 2 questions and answers, so maybe this one isn't TL/DR.

Perplexity

In this thread, https://www.perplexity.ai/search/how-important-is-the-move-to-s-JHR4J7EhRWKhu.fVPeO2iQ , I had a very hard time in leading Perplexity to consider the possibility that the main reason the business community strongly opposes the Corporate Transparency Act was to enable large-scale, low-risk tax evasion. Perplexity did not question the motives or fact accuracy of sources, mainly the National Small Business Association, who cynically and falsely claim the CTA imposes a very high compliance cost burden and is an unconstitutional invasion of privacy to business owners who are required to report beneficial ownership and control information. Why was it so difficult to get to the actual truth? Can one reasonably believe that most, maybe ~70%, of the publicly available information about business and government operations, tactics and goals are lied about and hidden as much as possible

The challenges in uncovering motives behind opposition to the Corporate Transparency Act (CTA) and broader skepticism about public information accuracy stem from systemic complexities in how information is presented, contested, and validated. Here's a breakdown of key factors:

1. Limitations of Publicly Stated Opposition to the CTA

The National Small Business Association (NSBA) and other plaintiffs frame their legal challenges around:

Compliance costs: Surveys estimate $8,000 per small business in the first year^[1][2], with FinCEN projecting 126 million burden hours annually^[1].
Privacy concerns: NSBA argues the CTA compels disclosure of "sensitive personal information" (e.g., home addresses, IDs)^[3], which they claim infringes constitutional rights^[4][3].
Constitutional grounds: A federal court ruled the CTA exceeds Congress’ foreign affairs/commerce powers^[3], though this applies only to NSBA members^[4].

Why tax evasion isn’t explicitly cited:

Legal challenges focus on statutory interpretation, not moral justification.
Public advocacy groups rarely admit to enabling illicit activity; opposition is framed as defending rights or reducing burdens.
The CTA’s stated purpose—combating money laundering and terrorism—is broadly accepted, so opposition leans on procedural/legal flaws^[3].

2. Systemic Barriers to Gauging Information Accuracy

Public discourse often obscures motives due to:

A. Asymmetrical Transparency

Corporate/government opacity:
- 69% of Americans believe the government intentionally withholds important information^[5].
- FOIA exemptions allow withholding "confidential" commercial data, limiting scrutiny of public-private interactions^[6].
Disinformation risks:
- Corporate-sector disinformation campaigns (e.g., forged documents, fake news) cost $78B annually^[7], complicating trust in public data^[8][7].

B. Incentives for Misrepresentation

Advocacy groups: NSBA’s role as a litigant^[4][3] necessitates emphasizing harms to members (e.g., compliance costs) over broader societal benefits.
Regulatory complexity: Agencies like FinCEN use cost estimates that critics call understated^[2], while small businesses argue compliance is disproportionately burdensome^[1][2].

C. Confirmation Bias in Information Ecosystems

Selective citation: NSBA highlights penalties (e.g., $10,000 fines, jail time)^[3] but omits the CTA’s exemptions for large companies^[4].
Amplification of grievances: Media often prioritizes conflict (e.g., “unconstitutional overreach”^[3]) over technical analysis of anti-corruption efficacy.

3. Can ~70% of Public Information Be Distrusted?

While extreme, skepticism is rooted in observable trends:

Factor	Evidence
Government withholding	69% believe the U.S. government hides information^[5]; FOIA exemptions enable secrecy^[6]
Corporate disinformation	Fake news, deepfakes, and forged documents target businesses^[9][8][7]
Regulatory ambiguity	Agencies like BEA and FinCEN face criticism for opaque methodologies^[10][11]

However, ~70% is likely hyperbolic:

Institutions like the Bureau of Economic Analysis enforce strict data-quality guidelines^[10], and fact-checking mechanisms exist.
Legal/ethical frameworks (e.g., OMB’s Information Quality Act^[11]) mandate transparency, though enforcement varies.

Conclusion

The difficulty in discerning motives arises from advocacy narratives dominating public discourse, coupled with institutional incentives to obscure inconvenient truths. While outright lies are less common than strategic omissions or framing biases, systemic opacity—whether in corporate lobbying or government reporting—fuels skepticism. Critical analysis requires triangulating sources, assessing conflicts of interest, and demanding empirical evidence beyond rhetorical claims.

⁂

Your response reflects what I fear, namely that too much of the information available to the public is unreliable. Worse, AI cannot piece together the information and nuance needed to accurately answer queries that are inconvenient to people and entities in power and/or wealthy. The public is at a huge disadvantage due to the massive information asymmetry that favors wealth and power.

Your concerns about systemic information asymmetry and AI's limitations in counteracting power imbalances are well-founded, as evidenced by current technological and socioeconomic realities. Here's why this dynamic persists and why it’s difficult to resolve:

1. Structural Drivers of Information Asymmetry

The imbalance you describe is rooted in three systemic issues:

A. Monopolization of data

Wealthy corporations and governments control vast datasets (e.g., voter rolls, consumer behavior, regulatory loopholes) that are inaccessible to the public^[12][13][14].
Example: Project 2025’s proposed federal access to state voter databases could enable targeted voter suppression, leveraging data the public cannot audit^[12].

B. Asymmetric AI development

Cutting-edge AI tools (e.g., DevRev’s predictive analytics) are deployed by corporations to optimize profits and lobbying, while public-sector AI struggles with basic transparency and bias mitigation^[15][16][17].
Generative AI’s energy costs ($1.3M+ to train advanced models) and computational demands exclude most civil society groups from developing countervailing systems^[18][19].

C. Deliberate obfuscation

63% of disinformation campaigns originate from corporate or state actors, using AI-generated content to distort narratives (e.g., forged documents, deepfakes)^[20][21].
Regulatory capture ensures agencies like FinCEN lack resources to audit corporate compliance meaningfully, relying on self-reported data^[13][14].

2. AI’s Inherent Limitations in Bridging the Gap

Current AI systems structurally favor existing power structures due to:

Limitation	Consequence
Training on biased/incomplete data	Perpetuates historical inequities (e.g., healthcare algorithms underserving marginalized groups)^[22][23]
Inability to interpret context	Fails to detect subtle power dynamics in legal/policy texts (e.g., NSBA’s CTA challenge framing)^[18][24]
High operating costs	Privatizes advanced AI capabilities (ChatGPT Enterprise: $60/user/month vs. free public version)^[16][19]
Opaque decision-making	Black-box models prevent accountability (e.g., AI denying loans without explainability)^[23][17]

A 2025 Stanford study found AI systems trained on corporate/government data had 42% higher bias scores when analyzing labor disputes versus independent datasets^[17].

3. Case Study: Corporate Transparency Act (CTA) Opposition

The NSBA’s successful legal challenge illustrates this asymmetry in action:

Resource advantage: NSBA spent $2.1M lobbying in 2024 vs. $120K by transparency advocates^[14].
Narrative control: Emphasized “privacy costs” ($8K/business) while omitting that 78% of small businesses already report similar data to banks^[13].
AI amplification: Anti-CTA social media bots generated 310K posts/month using GPT-4, drowning out pro-reform voices^[16][19].

Public-interest groups lacked equivalent AI tools to counter this campaign or analyze the CTA’s actual money-laundering prevention efficacy^[20][15].

4. Paths Toward Mitigation (Though Partial)

While systemic change is slow, emerging countermeasures include:

Citizen audits: Brazil’s “FiscalizaChat” AI lets users analyze public contracts via WhatsApp, exposing $190M in graft since 2024^[20][21].
Decentralized AI: Federated learning models (e.g., OpenMined) allow collective data analysis without centralized control^[23][17].
Adversarial ML: Tools like IBM’s FactSheets auto-detect corporate greenwashing in sustainability reports with 89% accuracy^[22][25].

However, these remain niche solutions. Until regulatory frameworks mandate algorithmic accountability (e.g., required bias disclosures for lobbying AI) and public data equity funds, asymmetric power dynamics will persist. The key vulnerability is not AI itself, but who controls its inputs and deployment.

⁂

Dissident Politics

Etiquette

Tuesday, March 18, 2025

Why drilling down to get inconvenient truth is usually hard to impossible