"AI trading" has been attached to enough scams that the phrase itself now tells you nothing. It sits on legitimate research desks and on social-media accounts promising to double your money by Friday. So the question "is AI trading a scam" has no clean yes-or-no answer. The useful question is narrower: how do you tell a real offer from a sold one?
The tests below are the same ones the CFTC and SEC published after a wave of AI-branded fraud. They are not about the technology. They are about what the seller lets you check.
AI trading is not inherently a scam, and it is not inherently anything. "AI" is a label, and a label cannot be honest or dishonest by itself. What can be judged is the offer wrapped around it: whether you can reproduce the results, whether the losing side is shown, whether the numbers arrive with an account size and a period, and whether anyone is promising an outcome. A real offer survives those questions. A sold one changes the subject.
Do not evaluate the technology, evaluate what you can verify. A trustworthy offer hands you a result you can reproduce and a downside stated out loud. A scam hands you a screenshot and a promise. The word "AI" is irrelevant to which one you are looking at.
This is the single most powerful test, and almost nothing fraudulent passes it. A real system gives you a result you can regenerate yourself, not a screenshot you have to believe. For a NinjaTrader 8 strategy, that means loading it into the Strategy Analyzer, running the same period, and seeing the same report print on your own machine. If the only evidence is an image of an equity curve or a statement you cannot recreate, treat it as marketing, not proof. A fabricated track record is easy to spot once you know the tells, covered in how to spot a fake track record.
Every real strategy loses. It has losing trades, losing streaks and a maximum drawdown, and an honest seller shows all three without being asked. A scam shows only the upside: the smooth curve, the big month, the win rate. The published Algomyth engine backtest, for example, states a worst drawdown of $5,273 on a mid-size account, and says plainly that its target sits at three times its risk, so break-even is a 25% win rate rather than a coin flip. Those are not flattering numbers, and that is the point. When a seller volunteers the ugly figures, they are more likely to be telling the truth about the pretty ones. What drawdown is and why it matters is in what is drawdown.
A profit figure means nothing without four things attached, and their absence is a red flag on its own.
The full worked version of why a bare number is worthless is in how much a trading bot actually makes.
An account size, a period, a drawdown, and a report you can reproduce yourself. Check the published numbers against every test on this page.
This is the fastest disqualifier. "Guaranteed returns", "risk-free", "passive income machine", a monthly percentage you can count on: these phrases are not optimism, they are the language regulators specifically warn about. The CFTC and SEC both issued advisories after a wave of AI-branded fraud, and the common thread was the promise, not the technology. Nobody who understands markets guarantees a return, because nobody can. An honest offer tells you it can lose, and tells you how much.
Machine learning has genuine, narrow uses in trading: classifying market conditions, filtering noise, processing large datasets faster than a person can. Those are real and unglamorous. What is not real is the claim that an AI has found a reliable way to predict price that everyone else missed. If such a pattern existed and were stable, it would be arbitraged away by firms with vastly more data and computing power than any retail product. The broader subject, and how much of it is genuine, is in the AI trading guide, and the same skepticism applied to "free" offers is in the truth about free trading bots.
A useful rule: prefer a system whose logic you can read and reproduce over one that hides behind "proprietary AI". A rule-based strategy can be inspected, backtested and understood. A black box asks for faith, and faith is exactly what a scam needs from you.
Hypothetical performance. The $5,273 drawdown and any other figure here comes from a backtest of the Algomyth engine over historical data, not a live account. Simulated results carry no financial risk and cannot fully reflect real execution. Past performance, real or simulated, does not guarantee future results.
You do not need to be technical to apply all four. Ask: can I reproduce this myself, or only look at it? Is the drawdown shown as plainly as the profit? Does every number come with an account size, a contract count, a period and a backtest-or-live label? Is anyone promising an outcome? A real offer answers the first three with a yes and the fourth with a firm no. A sold one gets uncomfortable and starts talking about the technology instead. The red flags to watch alongside these are collected in trading bot scams and their red flags.
No. AI trading is not inherently a scam or inherently legitimate, because "AI" is only a label and a label cannot be honest by itself. What determines whether an offer is trustworthy is everything wrapped around the label: whether you can reproduce the results, whether the downside is shown, whether numbers come with context, and whether anyone is promising an outcome. Judge the offer by what you can verify, and the technology it name-drops becomes almost irrelevant.
Look for a promise. "Guaranteed returns", "risk-free", "passive income" or a dependable monthly percentage are the exact phrases regulators warn about, because no real strategy can promise an outcome. The moment an offer guarantees a result, you can stop evaluating it. An honest seller tells you the system can lose and shows you how much, which is the opposite of what a scam does.
Because a result you can regenerate yourself cannot be faked, while a screenshot can be made in a minute. A real NinjaTrader 8 strategy lets you load it, run the same historical period, and watch the same report appear on your own machine. If the only evidence offered is an image you have to take on trust, treat it as advertising rather than proof. Reproducibility is the difference between evidence and decoration.
Almost never in the way it is sold. Machine learning has legitimate, narrow uses like filtering noise or classifying conditions, but the claim that an AI has found a stable, secret way to predict prices does not hold up, because such a pattern would be competed away by firms with far more resources than any retail tool. Be especially skeptical of a black box that cannot explain its own decisions. A rule-based system you can read and test is more trustworthy than proprietary AI you have to believe in.
It is built to pass the same tests this page describes. The results are a record published in full with a live signal log on top, the maximum drawdown is stated as plainly as the profit, every figure comes with the account size, contract count and period, and nothing is guaranteed. Those are not marketing choices, they are the whole argument. A system whose only defense is a screenshot and a promise is one you should walk away from.
Everything we write comes back to one idea: verify, then decide. The live terminal exists for exactly that.
Unlock Algomyth Live