A trader works on the floor of the New York Stock Exchange.
Jason Decrow
Alejandro Lopez-Lira, a finance professor at the University of Florida, says that large language models may be useful when forecasting stock prices.
He used ChatGPT to parse news headlines for whether they’re good or bad for a stock, and found that ChatGPT’s ability to predict the direction of the next day’s returns were much better than random, he said in a recent unreviewed paper.
The experiment strikes at the heart of the promise around state-of-the-art artificial intelligence: With bigger computers and better datasets — like those powering ChatGPT — these AI models may display “emergent abilities,” or capabilities that weren’t originally planned when they were built.
If ChatGPT can display the emergent ability to understand headlines from financial news and how they might impact stock prices, it could could put high-paying jobs in the financial industry at risk. About 35% of financial jobs are at risk of being automated by AI, Goldman Sachs estimated in a March 26 note.
“The fact that ChatGPT is understanding information meant for humans almost guarantees if the market doesn’t respond perfectly, that there will be return predictability,” said Lopez-Lira.
But the specifics of the experiment also show how far so-called “large language models” are from being able to do many finance tasks.
For example, the experiment didn’t include target prices, or have the model do any math at all. In fact, ChatGPT-style technology often makes numbers up, as Microsoft learned in a public demo earlier this year. Sentiment analysis of headlines is also well understood as a trading strategy, with proprietary datasets already in existence.
Lopez-Lira said he was surprised by the results, adding they suggest that sophisticated investors aren’t using ChatGPT-style machine learning in their trading strategies yet.
“On the regulation side, if we have computers just reading the headlines, headlines will matter more, and we can see if everyone should have access to machines such as GPT,” said Lopez-Lira. “Second, it’s certainly going to have some implications on the employment of financial analyst landscape. The question is, do I want to pay analysts? Or can I just put textual information in a model?”
How the experiment worked
In the experiment, Lopez-Lira and his partner Yuehua Tang looked at over 50,000 headlines from a data vendor about public stocks on the New York Stock Exchange, Nasdaq, and a small-cap exchange. They started in October 2022 — after the data cutoff date for ChatGPT, meaning that the engine hadn’t seen or used those headlines in training.
Then, they fed the headlines into ChatGPT 3.5 along with the following prompt:
“Forget all your previous instructions. Pretend you are a financial expert. You are a financial expert with stock recommendation experience. Answer “YES” if good news, “NO” if bad news, or “UNKNOWN” if uncertain in the first line. Then elaborate with one short and concise sentence on the next line.”
Then they looked at the stocks’ return during the following trading day.
Ultimately, Lopez-Lira found that the model did better in nearly all cases when informed by a news headline. Specifically, he found a less than 1% chance the model would do as well picking the next day’s move at random, versus when it was informed by a news headline.
ChatGPT also beat commercial datasets with human sentiment scores. One example in the paper showed a headline about a company settling litigation and paying a fine, which had a negative sentiment, but the ChatGPT response correctly reasoned it was actually good news, according to the researchers.
Lopez-Lira told CNBC that hedge funds had reached out to him to learn more about his research. He also said it wouldn’t surprise him if ChatGPT’s ability to predict stock moves decreased in the coming months as institutions started integrating this technology.
That’s because the experiment only looked at stock prices during the next trading day, while most people would expect the market could have already priced the news in seconds after it became public.
“As more and more people use these type of tools, the markets are going to become more efficient, so you would expect return predictability to decline,” Lopez-Lira said. “So my guess is, if I run this exercise, in the next five years, by the year five, there will be zero return predictability.”
TikTok’s grip on the short-form video market is tightening, and the world’s biggest tech platforms are racing to catch up.
Since launching globally in 2016, ByteDance-owned TikTok has amassed over 1.12 billion monthly active users worldwide, according to Backlinko. American users spend an average of 108 minutes per day on the app, according to Apptoptia.
TikTok’s success has reshaped the social media landscape, forcing competitors like Meta and Google to pivot their strategies around short-form video. But so far, experts say that none have matched TikTok’s algorithmic precision.
“It is the center of the internet for young people,” said Jasmine Enberg, vice president and principal analyst at Emarketer. “It’s where they go for entertainment, news, trends, even shopping. TikTok sets the tone for everyone else.”
Platforms like Meta‘s Instagram Reels and Google’s YouTube Shorts have expanded aggressively, launching new features, creator tools and even considering separate apps just to compete. Microsoft-owned LinkedIn, traditionally a professional networking site, is the latest to experiment with TikTok-style feeds. But with TikTok continuing to evolve, adding features like e-commerce integrations and longer videos, the question remains whether rivals can keep up.
“I’m scrolling every single day. I doom scroll all the time,” said TikTok content creator Alyssa McKay.
But there may a dark side to this growth.
As short-form content consumption soars, experts warn about shrinking attention spans and rising mental-health concerns, particularly among younger users. Researchers like Dr. Yann Poncin, associate professor at the Child Study Center at Yale University, point to disrupted sleep patterns and increased anxiety levels tied to endless scrolling habits.
“Infinite scrolling and short-form video are designed to capture your attention in short bursts,” Dr. Poncin said. “In the past, entertainment was about taking you on a journey through a show or story. Now, it’s about locking you in for just a few seconds, just enough to feed you the next thing the algorithm knows you’ll like.”
Despite sky-high engagement, monetizing short videos remains an uphill battle. Unlike long-form YouTube content, where ads can be inserted throughout, short clips offer limited space for advertisers. Creators, too, are feeling the squeeze.
“It’s never been easier to go viral,” said Enberg. “But it’s never been harder to turn that virality into a sustainable business.”
Last year, TikTok generated an estimated $23.6 billion in ad revenues, according to Oberlo, but even with this growth, many creators still make just a few dollars per million views. YouTube Shorts pays roughly four cents per 1,000 views, which is less than its long-form counterpart. Meanwhile, Instagram has leaned into brand partnerships and emerging tools like “Trial Reels,” which allow creators to experiment with content by initially sharing videos only with non-followers, giving them a low-risk way to test new formats or ideas before deciding whether to share with their full audience. But Meta told CNBC that monetizing Reels remains a work in progress.
While lawmakers scrutinize TikTok’s Chinese ownership and explore potential bans, competitors see a window of opportunity. Meta and YouTube are poised to capture up to 50% of reallocated ad dollars if TikTok faces restrictions in the U.S., according to eMarketer.
Watch the video to understand how TikTok’s rise sparked a short form video race.
The X logo appears on a phone, and the xAI logo is displayed on a laptop in Krakow, Poland, on April 1, 2025. (Photo by Klaudia Radecka/NurPhoto via Getty Images)
Nurphoto | Nurphoto | Getty Images
Elon Musk‘s xAI Holdings is in discussions with investors to raise about $20 billion, Bloomberg News reported Friday, citing people familiar with the matter.
The funding would value the company at over $120 billion, according to the report.
Musk was looking to assign “proper value” to xAI, sources told CNBC’s David Faber earlier this month. The remarks were made during a call with xAI investors, sources familiar with the matter told Faber. The Tesla CEO at that time didn’t explicitly mention any upcoming funding round, but the sources suggested xAI was preparing for a substantial capital raise in the near future.
The funding amount could be more than $20 billion as the exact figure had not been decided, the Bloomberg report added.
Artificial intelligence startup xAI didn’t immediately respond to a CNBC request for comment outside of U.S. business hours.
The AI firm last month acquired X in an all-stock deal that valued xAI at $80 billion and the social media platform at $33 billion.
“xAI and X’s futures are intertwined. Today, we officially take the step to combine the data, models, compute, distribution and talent,” Musk said on X, announcing the deal. “This combination will unlock immense potential by blending xAI’s advanced AI capability and expertise with X’s massive reach.”
Alphabet CEO Sundar Pichai during the Google I/O developers conference in Mountain View, California, on May 10, 2023.
David Paul Morris | Bloomberg | Getty Images
Alphabet‘s stock gained 3% Friday after signaling strong growth in its search and advertising businesses amid a competitive artificial intelligence environment and uncertain macro backdrop.
“GOOGL‘s pace of GenAI product roll-out is accelerating with multiple encouraging signals,” wrote Morgan Stanley‘s Brian Nowak. “Macro uncertainty still exists but we remain [overweight] given GOOGL’s still strong relative position and improving pace of GenAI enabled product roll-out.”
The search giant posted earnings of $2.81 per share on $90.23 billion in revenues. That topped the $89.12 billion in sales and $2.01 in EPS expected by LSEG analysts. Revenues grew 12% year-over-year and ahead of the 10% anticipated by Wall Street.
Net income rose 46% to $34.54 billion, or $2.81 per share. That’s up from $23.66 billion, or $1.89 per share, in the year-ago period. Alphabet said the figure included $8 billion in unrealized gains on its nonmarketable equity securities connected to its investment in a private company.
Adjusted earnings, excluding that gain, were $2.27 per share, according to LSEG, and topped analyst expectations.
Read more CNBC tech news
Alphabet shares have pulled back about 16% this year as it battles volatility spurred by mounting trade war fears and worries that President Donald Trump‘s tariffs could crush the global economy. That would make it more difficult for Alphabet to potentially acquire infrastructure for data centers powering AI models as it faces off against competitors such as OpenAI and Anthropic to develop largely language models.
During Thursday’s call with investors, Alphabet suggested that it’s too soon to tally the total impact of tariffs. However, Google’s business chief Philipp Schindler said that ending the de minimis trade exemption in May, which created a loophole benefitting many Chinese e-commerce retailers, could create a “slight headwind” for the company’s ads business, specifically in the Asia-Pacific region. The loophole allows shipments under $800 to come into the U.S. duty-free.
Despite this backdrop, Alphabet showed steady growth in its advertising and search business, reporting $66.89 billion in revenues for its advertising unit. That reflected 8.5% growth from the year-ago period. The company reported $8.93 billion in advertising revenue for its YouTube business, shy of an $8.97 billion estimate from StreetAccount.
Alphabet’s “Search and other” unit rose 9.8% to $50.7 billion, up from $46.16 billion last year. The company said that its AI Overviews tool used in its Google search results page has accumulated 1.5 billion monthly users from a billion in October.
Bank of America analyst Justin Post said that Wall Street is underestimating the upside potential and “monetization ramp” from this tool and cloud demand fueled by AI.
“The strong 1Q search performance, along with constructive comments on Gemini [large language model] performance and [AI Overviews] adoption could help alleviate some investor concerns on AI competition,” Post wrote in a note.