Connect with us

Published

on

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.”

Continue Reading

Technology

Epic touts new AI tools for patients and doctors at company’s annual meeting

Published

on

By

Epic touts new AI tools for patients and doctors at company's annual meeting

Visitors attend UGM 2025.

Courtesy of Epic

Space travelers, robots and, of course, artificial intelligence.

They were all on display on Tuesday at Epic Systems‘ annual Users Group Meeting, held at the health software giant’s 1,670-acre campus in Verona, Wisconsin.

Judy Faulkner, Epic’s 82-year-old CEO, dressed for the occasion in a purple wig with neon green shoes and an iridescent vest, reminiscent of the fictional character Buzz Lightyear from the “Toy Story” franchise.

At the science fiction-themed event, Faulkner told the crowd that Epic has roughly 200 different AI features in development that aim to assist patients, clinicians and insurers.

“We are combining the intelligence and curiosity of the human being with the investigative capabilities of gen AI,” Faulkner said, in front of thousands of health-care executives packed into an 11,400-seat underground auditorium.

Epic, one of the largest private technology companies in the country, is best known for its electronic health record, or EHR, software. An EHR is a digital version of a patient’s medical history that’s updated by doctors and nurses, and the technology is integral to the modern U.S. health-care system.

Epic’s software, which competes with Oracle Health (formerly Cerner), is used by 280 million Americans, according to the company. Many patients know of Epic because of its user portal called MyChart.

Last week, Epic announced MyChart Central, which will allow patients to log in to MyChart with just one set of credentials, rather than needing a username and password for each health system they visit. It’s equally helpful for health-care organizations, Faulkner said.

“You’ll spend less time handling patient calls and resetting passwords,” she said in her keynote on Tuesday. “Demographic changes like address need to be added only once.”

A new addition to the MyChart portal is the always-on Emmie assistant, which the company said will be able to answer questions about lab results, propose appointment times and suggest relevant screenings that patients can discuss with their doctor.

During Epic’s three-hour presentation, Faulkner and other executives introduced Emmie as well as other AI assistants the company calls Art and Penny, highlighting new capabilities that are coming in the next year and beyond.

Health-care executives attend UGM 2025.

Courtesy of Epic

The Art assistant is intended for clinicians, and is meant to act as an active AI digital colleague, the company said. Art will be able to anticipate information that a doctor might need, for instance, and can pull up information like blood pressure trends, update a patient’s family history and place orders.

The company also said Art will be able to draft clinical notes, which was one of the most highly anticipated announcements ahead of the conference. AI-powered clinical documentation tools, which are often called AI scribes, can take notes on patient visits in real time as doctors record their encounters, with a patient’s consent.

AI scribes have exploded in popularity as health-care executives search for solutions to help reduce staff burnout and daunting administrative workloads. Some startups in the space, including Abridge and Ambience Healthcare, have raised hundreds of millions of dollars from investors.

Epic said its AI charting tool is being built in collaboration with Microsoft. Epic and Microsoft have been working closely together for roughly two decades, and Microsoft’s DAX Copilot product is already a popular offering within the AI scribing market.

“We’re proud to be collaborating with Epic to explore how we can bring our core Dragon ambient AI technology to Epic’s new AI Charting capability to further improve care delivery,” Joe Petro, corporate vice president of Microsoft Health & Life Sciences said in a statement.

Epic’s Penny assistant is designed to help with revenue cycle management and other administrative needs, such as generating appeal letters for insurance claims that get denied. It can also help speed up medical coding by serving up suggestions, Faulkner said. Those two features are already live.

“With all the challenges health-care organizations are facing, we need to make sure our clinicians and our organizations are strong and doing well in order to be able to take care of patients,” Faulkner said.

Epic closed out its executive address by teasing new AI capabilities that are coming to Cosmos, which is a deidentified patient dataset clinicians can use to conduct research. Health systems have to opt-in to participate in Cosmos, and the database currently includes information from more than 1,760 hospitals and 300 million patients.

Epic said it’s building a set of proprietary foundation models, called Cosmos AI, based on this data. The company is still evaluating different applications of the models, and launched the Cosmos AI Lab to help researchers and data scientists learn more.

Executives said the models could be used to predict a timeline of a patient’s potential medical events, like whether they’re a readmission risk or could eventually experience a heart attack.

“We’re finding that it continues to improve as it sees more patients,” said Seth Hain, a senior vice president of research and development at Epic. “Having only used 8 billion encounters so far, we’re just getting started.”

WATCH: The health care sector was overdue for a bounce

Basham: The health care sector was overdue for a bounce

Continue Reading

Technology

SoftBank Group shares plunge over 9% as Asian tech stocks decline

Published

on

By

SoftBank Group shares plunge over 9% as Asian tech stocks decline

The logo of Japanese company SoftBank Group is seen outside the company’s headquarters in Tokyo on January 22, 2025. 

Kazuhiro Nogi | Afp | Getty Images

Shares of SoftBank Group plunged as much as 9.17% Wednesday, as technology stocks in Asia declined, tracking losses in U.S. peers overnight.

The Japanese tech-focused investment firm saw shares drop for a second consecutive session, following its announcement of a $2 billion investment in Intel. Intel shares rose 6.97% to close at $25.31 Tuesday stateside.

Stock Chart IconStock chart icon

hide content

SoftBank Group shares

Other Japanese tech stocks also declined, with semiconductor giant Advantest falling as much as 6.27%. Meanwhile, shares in Renesas Electronics and Tokyo Electron were last seen trading 2.46% and 0.75% lower, respectively.

Technology companies in South Korea, Taiwan and Hong Kong, also fell after U.S. tech stocks dropped overnight spurred by declines in artificial intelligence darling Nvidia‘s shares.

U.S. Commerce Secretary Howard Lutnick is considering the federal government taking equity stakes in semiconductor companies that get funding under the CHIPS Act for building plants in the U.S, sources familiar with the matter told Reuters. The U.S. CHIPS and Science Act seeks to boost the country’s semiconductor industry, scientific research and innovation.

Shares of Taiwanese chip company TSMC and manufacturer Hon Hai Precision Industry — known globally as Foxconn — declined 1.69% and 2.16%, respectively. TSMC manufactures Nvidia’s high-performance graphics processing units that help power large language models, while Foxconn has a strategic partnership with Nvidia to build “AI factories.” 

Meanwhile, South Korean tech stocks mostly fell with shares of chipmaker SK Hynix down 3.33%. Samsung Electronics, however, rose 0.75%.

TSMC, Samsung and SK Hynix are among companies that have received funding under the CHIPS Act.

Over in Hong Kong, the Hang Seng Tech index lost 0.87% in early trade.

The worst performing stocks on the index were Kuaishou Technology which declined 4.8%, JD Health International which dropped 3.31% and Horizon Robotics which lost 2.29%.

Losses were also seen tech majors Alibaba Group, down 1.44%, and Xiaomi Corp which lost 1.34%.

Continue Reading

Technology

Palantir stock slumps 9%, falling for a fifth straight day from record

Published

on

By

Palantir stock slumps 9%, falling for a fifth straight day from record

CEO of Palantir Technologies Alex Karp attends the Pennsylvania Energy and Innovation Summit on the campus of Carnegie Mellon University in Pittsburgh, Pennsylvania on July 15, 2025.

Andrew Caballero-reynolds | Afp | Getty Images

Palantir‘s stock slumped more than 9% on Tuesday, falling for a fifth straight day to continue its pullback from all-time highs.

The artificial intelligence software provider’s stock has slid more than 15% over the last five trading sessions, after a stellar earnings report earlier this month propelled shares to all-time highs. The report was Palantir’s first-ever $1 billion revenue quarter.

Tuesday’s dip coincided with a broader market pullback.

Palantir is the most significant gainer to date in the S&P 500 in 2025, up more than 100%.

Read more CNBC tech news

Shares have more than doubled as the company benefits from ongoing AI enthusiasm, scooping up government contracts with President Donald Trump pushing to overhaul agencies.

Palantir’s ascent has pushed the company into a list of top 10 U.S. tech firms and 20 most valuable U.S. companies, while also making shares incredibly expensive to own. Its forward price-to-earnings ratio, which tracks future earnings relative to share price, has soared past 245 times.

By comparison, technology giants such as Microsoft and Apple carry a P/E of nearly 30 times and rake in significantly greater quarterly revenues. Meta‘s and Alphabet‘s P/E ratios hover in the 20s.

What to know about Palantir's engineer-led sales strategy

Continue Reading

Trending