AI-powered investment tools are rapidly gaining adoption among retail investors, with usage up 46% in 2025 and robo-advisor AUM reaching $1.8 trillion. While AI can reduce human biases like loss aversion and overconfidence, significant risks exist: black-box opacity makes decisions hard to explain or audit, herding behavior among similar algorithms can amplify volatility, and flash crash history shows how algorithmic cascades spiral out of control. Regulatory bodies including the SEC, CFTC, FCA, and ESMA are building guardrails, but fragmented approaches create compliance challenges. Key recommendations include mandatory explainability (XAI), independent auditing, stress testing against historical crises, bias reporting, and education integration so investors understand what AI can and cannot do. The conclusion is nuanced: AI can genuinely improve retail investor outcomes when implemented transparently, but risks systemic harm when deployed opaquely or without adequate oversight.