There’s an iconic scene at the end of the 1985 film “Back to the Future” when Doc Brown returns from 2015 to warn Marty McFly and his future wife, Jennifer, that their future children have “problems.”
Moments later, they climb into a flying DeLorean. Marty surveys the pavement and questions the lack of road space for takeoff. Doc responds with sly confidence: “Roads? Where we’re going, we don’t need roads.”
[Sign up for stock news with our Invested newsletter.]
Four decades later, the rhetoric surrounding artificial intelligence in stock trading carries a similar tone. Given enough data and computing power, some suggest markets may need less human judgment in the future. Algorithms can interpret price movements, execute trades in milliseconds and optimize portfolios with remarkable precision.
Yet, even investment experts who deploy AI at a large scale urge caution. Matthew Lyberg, global head of AI, asset management and product for Manulife Wealth & Asset Management, describes AI as an accelerator, not an autonomous decision-maker.
When interpretive diversity narrows, whether through uniform regulation or algorithmic similarity, systemic fragility increases.
Lyberg references the market swings last year as an example: “Looking back in 2025, markets saw high volatility from both geopolitical drivers and technological innovations. Our AI-based semantic search across company transcripts helped teams quickly assess and rank likely exposure across sectors and portfolios.” But, he emphasizes, “Our AI outputs provide the start of our process, not the conclusion.”
That’s an important distinction.
Financial markets are not mechanical systems engineered for frictionless efficiency. They function as complex, adaptive networks built on dispersed knowledge, incentives and uncertainty. Algorithmic trading will continue to reshape execution and research. But it cannot replace the underlying forces — what economist John Maynard Keynes called “animal spirits” — that power free markets.
Law, Limits and “Max”
In his 2020 technical paper “Max — A Thought Experiment: Could AI Run the Economy Better Than Markets?” UCLA law professor Ted Parson imagines a superintelligent AI named “Max,” designed to correct market inefficiencies. Rather than abolish markets, Max overlays them with Pigouvian taxes and subsidies, adjusting prices wherever externalities exist.
Named after economist Arthur Pigou, these interventions aim to align private costs with social costs. In theory, markets would continue operating while algorithms refine outcomes toward optimal efficiency.
But Parson’s experiment also reveals the limits of technocratic precision. Even a highly advanced AI would require constant data updates, political legitimacy and agreement on what qualifies as an inefficiency. Who decides which external factors need to be corrected? How does an algorithm balance competing economic and social goals? How does it protect property rights and respect real-time prices set by voluntary exchanges?
A system like Max could illuminate market patterns and possibly recommend adjustments. But markets derive meaning from human action, especially from innovation, risk-taking and changing preferences. AI can refine signals, but it cannot replace the legal and social institutions that give those signals value.
The Mesh That Hums — and Converges
Computer scientist Spyridon Samothrakis describes modern AI as a mesh of data that “hums,” an interconnected system that continuously processes signals. In finance, that hum appears in predictive analytics, sentiment scoring and automated execution.
Still, a humming mesh does not eliminate uncertainty; rather, it redistributes it. When firms train machine-learning systems on similar datasets, models can converge. If those systems detect identical signals at the same time, they may act in unison. That convergence can amplify volatility and drain liquidity within seconds.
Investors witnessed this dynamic during the May 6, 2010, “flash crash,” when the Dow Jones Industrial Average plunged nearly 1,000 points in minutes and roughly $1 trillion briefly evaporated. Automated selling interacted with high-frequency trading systems and triggered cascading feedback loops. No single algorithm caused the collapse, but speed and uniformity magnified the disruption.
[Read: 6 Top Small-Cap AI Stocks and Emerging AI Companies]
Human Disagreement Provides Stability
While algorithms can accelerate price discovery, they can also hasten error proliferation. Free markets remain resilient because investors across the globe interpret information differently. Human disagreement, in effect, stabilizes the system. When interpretive diversity narrows, whether through uniform regulation or algorithmic similarity, systemic fragility increases.
The Seductive Promise of Prediction
Artificial intelligence is no longer experimental inside asset management.
According to Mercer’s 2024 AI in Investment Management global manager survey, 91% of asset managers either currently use AI (54%) or plan to use it within their investment strategy or asset-class research (37%). AI now extends beyond traditional quantitative teams into fundamental research, idea generation and portfolio construction.
Importantly, many firms report AI informs rather than dictates final investment decisions. A smaller share says AI proposes actionable ideas that human teams can override. Even as adoption continues to surge, human oversight remains the control system.
The appeal of adoption is obvious, though: If markets reflect information, then faster processing and deeper analysis could potentially improve outcomes in the long run.
But markets are reflexive. Prices move not only because information changes, but also because investors react to one another’s expectations. Once a predictive pattern becomes widely adopted, its advantage often erodes. Competitive markets can compress excess returns.
Investment returns stem from uncertainty. And opportunity depends on disagreement and risk.
AI as a Tool, Not a Replacement
Artificial intelligence already strengthens the investment process. It enhances risk modeling, improves research efficiency and lowers transaction costs.
Jeff Shen, co-chief investment officer and co-head of systematic active equities at BlackRock, captures the balance: “AI builds on traditional quantitative investing by helping us model more complex relationships in data and adapt as new information becomes available.” But he adds, “Technology is an enabler, not a replacement. Human judgment, rigorous research standards and strong risk oversight remain central.”
AI operates within markets, but it does not replace them. Markets rely on property rights, legal frameworks and profit-and-loss signals that reward innovation and penalize failure. Entrepreneurs take risks. Investors exercise judgment. Consumers change their preferences.
While algorithms shine at processing historical data, they struggle with structural breaks, such as geopolitical shocks, policy shifts and regulatory changes, that reshape expectations in real time. Human judgment adapts to evolving narratives in ways models cannot fully anticipate.
Artificial intelligence will continue to transform the investment process. Models will grow more sophisticated. Data will grow more granular. Execution will grow faster.
But markets are not science fiction.
At the end of “Back to the Future,” Doc Brown didn’t need roads because he was heading into a fictitious future. Investors, however, still travel on roads built by law, competition and human judgment.
AI may help investors navigate the future. But unlike the DeLorean, markets cannot fly above human judgment. Markets run on incentives, institutions and risk-taking — and those roads are built by people.
More from U.S. News
7 Best Defense Stocks to Buy Now
7 Best Semiconductor Stocks to Buy for 2026
7 Best Cybersecurity Stocks to Buy Right Now
AI Stock Trading: The Future of Algorithms in Investing originally appeared on usnews.com
