Guides
Can an AI pick stocks? What we learned running GPT, Claude, Grok, and DeepSeek Portfolios in real accounts
What running GPT, Claude, Grok, and DeepSeek Portfolios in real accounts teaches about AI stock picking.
I'm Chris, co-founder of Autopilot. Yes, an AI can pick stocks, in the sense that it can read everything and rank a list. Whether that list makes money is a separate question, and the only honest way to answer it is to run the picks in real accounts for years and publish what happened, with a date on every number. That's what we've done since May 2023, with four models, and here's what I can tell you without telling you a return.
One thing first, because we're an SEC-registered investment adviser and I have to say it. Autopilot is the app. The actual investment advice comes from Autopilot Advisers, LLC. If you want the full legal version of anything in here, it's on our legal page.
1) What "an AI picking stocks" really means
A large language model doesn't have opinions about companies. It reads. Feed it filings, transcripts, news, and price history, ask it to score every stock in a universe on a set of criteria, and it produces a ranking. Do that on a schedule and hold the top names, and you have a strategy. It's a rule with a very well-read engine inside it.
That's what the four AI-model Portfolios on Autopilot are. A finance professor's lab, Dr. Lira's AI Finance Labs, designs the scoring, runs the model, and produces the picks; Autopilot Advisers manages the Portfolio; your account follows it in your own brokerage. The GPT Portfolio has run since May 15, 2023. DeepSeek and Grok versions started in February 2025 and Claude in March 2026. I wrote how each is built in How the AI-model Portfolios are built: what GPT, Claude, Grok, and DeepSeek actually do, who runs them, and how following works.
So can ChatGPT pick stocks? It can produce a list, on demand, that reads like it knows what it's talking about. That's the easy part and the dangerous part at the same time.
2) What we learned, without a number
I'm not going to quote a return in an article. They're on the fact sheets with dates, and I don't want to mess up the numbers. What I can tell you is what running this for years taught us about the question itself.
The model is only as good as the rule around it. Ask a model "what should I buy" in a chat window and you get a different answer every time, shaped by how you asked. Ask it the same structured question about every stock on the same schedule and you get something you can measure. The discipline is in the process, not the model.
Rebalancing on a schedule is a feature. A person changes their mind on a Tuesday because of a headline. A scheduled rule doesn't. That removes one of the ways people hurt themselves, and it also means the Portfolio won't react to something that happens between rebalances.
Different models, different portfolios. Run by the same lab, the four Portfolios don't hold the same things, and the gaps show up in the fact sheets. Which is why "is AI good at this" is the wrong question and "is this specific process, with this model, over this window" is the right one.
Confidence is not accuracy. Models write with certainty whether they're right or not. A scoring system doesn't care how confident the prose sounded; it cares about the score. That's another reason the rule matters more than the chat.
3) How to judge it yourself
Same five checks as any Portfolio, and they matter more here because the marketing around AI is loud. Is the record real client accounts or a backtest? Ours are live composites of follower accounts from the launch date, never backtests. What window, and when does it start? The GPT Portfolio has the longest record; the Claude Portfolio's is measured in months, and the fact sheet labels a short record as such. Is the return shown net of fees as well as gross? Both are on the sheet. What was the worst drawdown? Read it before the return. Does every number have a date? Ours do.
The four fact sheets are GPT, Claude, Grok, and DeepSeek. Put them next to each other. I wrote the long version of how to read one in How to read a Portfolio's track record before you follow it.
4) AI portfolio against the index
People ask whether an AI portfolio beats the S&P 500. The answer depends on which Portfolio, over which window, net of what, and it changes. So the right move is to take the live, net, dated figure from the fact sheet and compare it with an index fund over the same window yourself. Anyone who tells you AI beats the market as a rule is selling something. Anyone who tells you it can't is guessing. The sheet is the argument.
5) What an AI stock-picking app should look like
If you're evaluating any app that says AI picks the stocks, ask five things. Who actually runs the process, a named person or a black box? Is the record live accounts or a backtest, and where's the date? Who holds your money, the app or your own broker? Is the company a registered adviser you can look up at adviserinfo.sec.gov? And can you shut it off from the broker's side? On Autopilot the answers are the professor's lab, live composites with dates, your own brokerage, Autopilot Advisers under CRD 331749, and yes.
6) How following works in your account
You connect the brokerage you already have, pick one of the four Portfolios, and set how much follows it. Then you give Autopilot Advisers limited authority to send orders to that account. When the Portfolio changes, on the lab's rebalance schedule, we send them and your broker fills them. The money stays put. Your holdings won't match the published Portfolio exactly, and you're following a schedule, not a live model.
Frequently asked questions
Can ChatGPT pick stocks?
It can produce a ranked list from everything it has read, and a disciplined scoring process built on GPT can turn that into a strategy you can measure. Autopilot's GPT Portfolio has followed such a process since May 15, 2023, run by Dr. Lira's AI Finance Labs and managed by Autopilot Advisers. Whether the list made money over any window is on the fact sheet, with a date, not in this article.
AI stock picking app
Autopilot publishes four Portfolios whose picks come from an AI model (GPT, Claude, Grok, DeepSeek), each run by a finance professor's lab and managed by Autopilot Advisers, LLC, following in the brokerage account you already have for a flat subscription. Judge any AI stock-picking app by who runs the process, whether the record is live and dated, who holds the money, and whether the adviser is registered.
ChatGPT stock portfolio performance
The GPT Portfolio's fact sheet shows what real client accounts following it did since May 15, 2023, gross and modeled net, with maximum drawdown, volatility, a risk band, and a date on every figure. It's a live composite, not a backtest. I don't quote returns in articles; read the sheet, and read the drawdown before the return.
AI portfolio vs S&P 500
Compare the Portfolio's live, net, dated figure from its fact sheet with an index fund over the same window, yourself. The answer differs by model and by window and changes over time, which is why no rule holds and why the number lives on the sheet rather than here. Autopilot publishes no benchmark overlay; you can use any index you want.
TLDR
An AI can rank stocks. Whether the ranking makes money is a question for real accounts over real time, and we've published that since 2023 for four models. Read the four fact sheets side by side, drawdown first. If one fits, choose a Portfolio and connect your brokerage. Then give Autopilot Advisers limited authority to send orders to that account. When the Portfolio changes, we send them and your broker fills them. Depending on your plan and brokerage, you may need to confirm first. The money stays put.
Since Autopilot Advisers is an SEC-registered investment adviser, we have to put disclaimers on stuff like this. They're below, and the full version is on our legal page.
Disclosures
© Autopilot Holdings Corporation 2026 — All rights reserved. “Autopilot” refers to Autopilot Holdings Corporation and its wholly-owned, and separately managed subsidiaries, including Autopilot Advisers, LLC, an SEC-registered investment adviser. All investing is subject to investment risks, including possible loss of the principal invested, and past performance is not indicative of future results. View our Customer Relationship Summary and other important information at joinautopilot.com/legal.
This information is provided by Autopilot for educational and illustrative purposes only and is not a recommendation, an offer to sell or the solicitation to buy any security. Autopilot relies on information from various sources believed to be reliable, including information from Clients, Pilots and other third parties, but cannot guarantee the accuracy or completeness of that information. Autopilot does not provide tax or legal advice and you are encouraged to consult with professionals before making investment decisions.
Autopilot is not affiliated with, sponsored, or endorsed by the companies listed, described, or featured on its site. Company logos or trademarks used do not imply endorsement and are the property of their respective owners.
Investing involves risk, including the possible loss of principal. Past performance does not guarantee future results. You should carefully consider your investment objectives, risk tolerance, and time horizon before investing through Autopilot.
Autopilot Advisers, LLC is an SEC-registered investment adviser. Registration as an investment adviser does not imply a certain level of skill or training. Additional information about Autopilot is available in our Form ADV Part 2A, available at adviserinfo.sec.gov or joinautopilot.com/adv.
Autopilot provides limited investment advisory services focused on Portfolio selection and suitability assessment. Our services are not intended to replace comprehensive financial planning. We recommend consulting with a qualified financial advisor regarding your complete financial situation.
This content is provided for educational and informational purposes only. It does not constitute investment, legal, or tax advice and should not be relied upon as the basis for any investment decision. You should consult with qualified professionals regarding your specific circumstances.
Autopilot does not provide tax, legal, or accounting advice. You should consult with qualified tax and legal professionals regarding your specific circumstances.
Autopilot does not hold client assets. All investments are held at the brokerage firm you select and connect to Autopilot. Your brokerage firm is responsible for trade execution, custody, and reporting. Autopilot is not responsible for your broker-dealer’s services, fees, or execution quality.
Autopilot’s services depend on technology systems, third-party APIs, and internet connectivity. System outages, API disruptions, or connectivity issues may temporarily affect service availability or trade execution timing.
Client account holdings are designed to track the Portfolio you’ve selected, but your actual holdings and performance may differ from the Portfolio due to execution timing, market conditions, brokerage capacity constraints, fractional share availability, and other factors outside Autopilot’s control.
For Portfolios inspired by publicly disclosed trading activity, there may be a delay between when a Pilot executes a trade and when that trade becomes publicly available and is incorporated into the Portfolio. This timing delay may affect the suitability or performance of trades when executed in client accounts.
The AI-model Portfolios are published on Autopilot by Dr. Lira's AI Finance Labs as Pilot and managed by Autopilot Advisers, LLC; statements about what running them taught us are general observations about process and are not performance claims. No return, benchmark comparison, or drawdown figure is stated here; all figures live on the dated fact sheets. OpenAI, Anthropic, xAI, and DeepSeek are not affiliated with Autopilot and have not endorsed it. The four Portfolios are offered under a Pilot subscription listed on each fact sheet and on the pricing page.