Research On Criminal Responsibility In Ai-Assisted Algorithmic Financial Fraud, Automated Trading, And Market Manipulation
1. Michael Coscia – High-Frequency Trading Spoofing (USA, 2015)
Facts:
Michael Coscia used high-frequency trading algorithms to place large “fake” orders in commodity futures (gold, soybeans, copper, etc.) on U.S. exchanges.
His program would place orders to create the illusion of demand or supply, then cancel them before execution and profit from the market reaction.
Over a few months, this scheme earned him about $1.4 million.
Legal Issues:
Coscia exploited automated trading to manipulate market prices (spoofing).
Key question: Can a human be criminally liable when the algorithm does all the trading?
Court ruled: Yes. Liability falls on the human who designs and deploys the algorithm with intent to defraud or manipulate.
Outcome:
Convicted on 12 counts (6 counts of commodities fraud and 6 counts of spoofing).
Sentenced to 3 years in prison.
Implications:
Established a precedent that using automated systems for market manipulation is criminally actionable.
Highlights the need for oversight, logs, and monitoring of algorithmic trading systems.
2. Navinder Singh Sarao – Flash Crash Manipulation (USA, 2010/2015)
Facts:
Sarao used an automated trading program to manipulate E-mini S&P 500 futures.
Placed and canceled large orders rapidly to create a false market perception.
His activity was partly responsible for the “Flash Crash” of May 6, 2010, when the Dow dropped 600 points in minutes.
Legal Issues:
Algorithmic trading exploited market microstructure for manipulation.
Court emphasized that human intent to manipulate is sufficient for liability even if execution is automated.
Outcome:
Charged with commodity fraud and market manipulation.
Served 1 year in U.S. prison after extradition from the UK.
Implications:
Showed how automated systems can amplify market volatility.
Reinforced human accountability in algorithmic trading schemes.
3. Mina Tadrus – AI Hedge Fund Fraud (USA, 2025)
Facts:
Founded a hedge fund claiming to use AI-based algorithmic trading.
Collected over $5.7 million from investors, but most funds were misappropriated; little was invested in algorithmic trading.
Funds were used for personal expenses and paying earlier investors (Ponzi-like activity).
Legal Issues:
Fraudulent misrepresentation of AI trading capabilities.
Raises the question: Is claiming AI/algorithmic trading without substance a crime? Court said yes.
Outcome:
Pleaded guilty to wire fraud.
Sentenced to 30 months in prison and ordered to pay restitution.
Implications:
Misrepresentation of algorithmic or AI capabilities is actionable fraud.
Highlights investor risk in AI-labeled financial products.
4. Jian Wu – Manipulation of Algorithmic Trading Models (USA, 2025)
Facts:
Wu, a quantitative researcher, manipulated 14 algorithmic trading models in a hedge fund.
Altered models to appear predictive and unique while generating large losses for clients (~$165 million).
He profited from performance-linked compensation (~$23.5 million).
Legal Issues:
Manipulating algorithmic models constitutes securities and wire fraud.
Raises questions about internal accountability and algorithmic governance.
Outcome:
Charged with securities and wire fraud; case ongoing.
SEC filed parallel civil charges.
Implications:
Shows the need for strong model governance, audits, and oversight.
Highlights criminal responsibility for individuals altering algorithmic systems.
5. Xiaosong Wang & Jiali Wang – Market Manipulation via Multiple Accounts (USA, 2022-2024)
Facts:
Used multiple brokerage accounts to artificially inflate/deflate prices of thinly traded stocks.
Placed small orders to signal demand or supply, executed large trades in the opposite direction, then canceled initial orders.
Illicit gains reached millions.
Legal Issues:
Market manipulation using automated/semi-automated techniques, though not full AI.
Demonstrates that law applies to algorithm-style manipulation even if manual execution is involved.
Outcome:
Pleaded guilty; forfeited illicit proceeds.
Strengthened enforcement against layering/canceling schemes.
Implications:
Confirms human responsibility for market manipulation even with automated trading techniques.
Shows the regulatory importance of monitoring account activity and algorithmic trading patterns.
Key Takeaways from These Cases
Human Responsibility: Liability rests with humans controlling or designing the algorithm.
Algorithmic Amplification: Automated systems can execute fraudulent trades at high speed and scale.
AI/Algorithm Claims: Misrepresentation of algorithmic capabilities is fraud.
Internal Governance: Strong oversight, audits, and model verification are essential.
Regulatory Scope: Existing laws (spoofing, fraud, market manipulation) extend to algorithmic/AI-assisted contexts.
RELATED Blog
comments