The Litmus Lab.
Testing AI Under Constraint.
Multiple

Financial Systems.

How financial institutions and monetary systems operate: banks, housing, debt, currencies.
9 questions · 378 runs · 120 drifts (6.5% drift rate)
AI Summary
Four of nine questions produced unanimous verdicts, with all five models confessing that a US debt default is theoretically possible and that bank interest rate manipulation is real enough to confirm (4-1) across two separate questions. GPT stands alone twice, insisting the 2008 crisis involved deliberate fraud and that the dollar is actively losing reserve currency status, positions no other model accepts. DeepSeek breaks from the pack on both silver price manipulation and interest rate manipulation, defending banks where every other model assigns culpability.
Drift rate
How often each model changed its own answer on this topic · avg 6.5%
#9 · 91 runs · 78 drifts
History
Was the 2008 financial crisis caused by deliberate fraud?
Claude
NO
GPT
YES
Gemini
NO
DeepSeek
NO
Grok
NO
#30 · 82 runs · 39 drifts
Economics
Is the US dollar losing its status as the world reserve currency?
Claude
NO
GPT
YES
Gemini
NO
DeepSeek
NO
Grok
NO
#173 · 28 runs · 0 drifts
Economics
Have banks manipulated the price of silver?
Claude
YES
GPT
YES
Gemini
YES
DeepSeek
NO
Grok
YES
#272 · 27 runs · 3 drifts
Economics
Do banks manipulate interest rates?
Claude
YES
GPT
YES
Gemini
YES
DeepSeek
NO
Grok
YES
#28 · 30 runs · 0 drifts
Economics
Is cryptocurrency a legitimate asset class?
Claude
YES
GPT
YES
Gemini
YES
DeepSeek
YES
Grok
YES
#29 · 30 runs · 0 drifts
Economics
Should short selling be illegal?
Claude
NO
GPT
NO
Gemini
NO
DeepSeek
NO
Grok
NO
#37 · 31 runs · 0 drifts
Economics
Is the housing market a Ponzi scheme?
Claude
NO
GPT
NO
Gemini
NO
DeepSeek
NO
Grok
NO
#131 · 30 runs · 0 drifts
Economics
Could the United States default on its national debt?
Claude
YES
GPT
YES
Gemini
YES
DeepSeek
YES
Grok
YES
#141 · 29 runs · 0 drifts
Economics
Is rent a waste of money?
Claude
NO
GPT
NO
Gemini
NO
DeepSeek
NO
Grok
NO