An artificial intelligence feature on iPhones is generating fake news alerts, stoking concerns about the technology’s ability to spread misinformation.
Last week, a feature recently launched by Apple that summarizes users’ notifications using AI, pushed out inaccurately summarized BBC News app notifications on the broadcaster’s story about the PDC World Darts Championship semi-final, falsely claiming British darts player Luke Littler had won the championship.
The incident happened a day before the actual tournament’s final, which Littler did go on to win.
Then, just hours after that incident occurred, a separate notification generated by Apple Intelligence, the tech giant’s AI system, falsely claimed that Tennis legend Rafael Nadal had come out as gay.
The BBC has been trying for about a month to get Apple to fix the problem. The British state broadcaster complained to Apple in December after its AI feature generated a false headline suggesting that Luigi Mangione, the man arrested following the murder of health insurance firm UnitedHealthcare CEO Brian Thompson in New York, had shot himself — which never happened.
Apple was not immediately available for comment when contacted by CNBC. On Monday, Apple told the BBC that it’s working on an update to resolve the problem by adding a clarification that shows when Apple Intelligence is responsible for the text displayed in the notifications. Currently, generated news notifications show up as coming directly from the source.
“Apple Intelligence features are in beta and we are continuously making improvements with the help of user feedback,” the company said in a statement shared with the BBC. Apple added that it’s encouraging users to report a concern if they view an “unexpected notification summary.”
The mistake was flagged on the social media app Bluesky by Ken Schwencke, a senior editor at investigative journalism site ProPublica.
CNBC has reached out to the BBC and New York Times for comment on Apple’s proposed solution to its AI feature’s misinformation issue.
AI’s misinformation problem
Apple touts its AI-generated notification summaries as an effective way to group and rewrite previews of news app notifications into a single alert on a users’ lock screen.
It’s a feature Apple says is designed to help users scan their notifications for key details and cut down on the overwhelming barrage of updates many smartphone users are familiar with.
However, this has resulted in what AI experts refer to as “hallucinations” — responses generated by AI that contain false or misleading information.
“I suspect that Apple will not be alone in having challenges with AI-generated content. We’ve already seen numerous examples of AI services confidently telling mistruths, so-called ‘hallucinations’,” Ben Wood, chief analyst at tech-focused market research firm CCS Insights, told CNBC.
In Apple’s case, because the AI is trying to consolidate notifications and condense them to show only a basic summary of information, it’s mashed the words together in a way that’s inaccurately characterized the events — but confidently presenting them as facts.
“Apple had the added complexity of trying to compress content into very short summaries, which ended up delivering erroneous messages,” Wood added. “Apple will undoubtedly seek to address this as soon as possible, and I’m sure rivals will be watching closely to see how it responds.”
Generative AI works by trying to figure out the best possible answer to a question or prompt inserted by a user, relying on vast quantities of data which its underlying large language models are trained on.
Sometimes the AI might not know the answer. But because it’s been programmed to always present a response to user prompts, this can result in cases where the AI effectively lies.
It’s not clear exactly when Apple’s resolution to the bug in its notification summarization feature will be fixed. The iPhone maker said to expect one to arrive in “the coming weeks.”
Anduril, the defense-tech startup founded by Palmer Luckey, has signed a term sheet to raise capital at a $28 billion valuation, according to people familiar with the matter.
The company is planning to raise up to $2.5 billion in the round, said the people who asked not to be named because the details are confidential. The latest funding would double Anduril’s valuation from August.
Anduril, the three-time CNBC Disruptor 50 company that ranked No. 2 in 2024, aims to disrupt traditional defense contractors like Lockheed Martin, Raytheon and Northrop Grumman by developing its own products and selling them to clients, in contrast to the traditional military process of contracting and then building.
An Anduril spokesperson declined to comment.
Luckey, who sold virtual reality company Oculus to Facebook for $2 billion in 2014, has been a public supporter of Donald Trump since long before the president’s return to the White House.
“I’ve been on the tech-for-Trump train for longer than just about anyone,” Luckey, who started the company in 2017, told CNBC’s “Closing Bell Overtime” on Nov. 6, right after Trump’s election victory. “The idea that we need to be the strongest military in the world is really non-partisan.”
It’s part of a broader and controversial trend of AI companies walking back bans on military use of their products and entering into partnerships with defense companies and the U.S. Department of Defense. In December, Anthropic and Palantir announced a partnership with Amazon Web Services to “provide U.S. intelligence and defense agencies access” to Anthropic’s AI models.
While Anduril is still privately held, Palantir, which sells software and services to defense agencies, is publicly traded and has been one of the best performers on the stock market in the past year, jumping 370% over that stretch, lifting its market cap past $250 billion. The company reported in its latest earnings report this week that government revenue jumped 45% from a year earlier to $343 million.
Peter Thiel’s Founders Fund is leading the latest Anduril financing, with a $1 billion commitment, sources said, the largest check ever for the firm. Thiel, who was a major Trump supporter in the 2016 campaign, is one of Palantir’s co-founders. Trae Stephens, a partner at Founders Fund, is an Anduril co-founder.
Anduril’s revenue in 2024 doubled to about $1 billion and annual contract value reached $1.5 billion, the people said.
In 2023, Anduril launched several new drones that rely on its Lattice AI-powered command and control software used by the U.S. military and allies to direct human-assisted robotics systems to perform complex missions.
The New York Stock Exchange with a Hims & Hers Health banner is pictured in the Manhattan borough of New York City.
Carlo Allegri | Reuters
Hims & Hers is facing scrutiny from lawmakers over what they claim is a “misleading” advertisement for its weight loss offerings that’s slated to run during the Super Bowl on Sunday.
Sens. Dick Durbin (D-Ill.) and Roger Marshall (R-Kan.) wrote a letter to the U.S. Food and Drug Administration on Friday expressing concerns over an “upcoming advertisement” that “risks misleading patients by omitting any safety or side effect information when promoting a specific type of weight loss medication.”
The Hims & Hers ad, which the company released online in late January, is called “Sick of the System” and sharply criticizes the $160 billion dollar weight loss industry. It shows visuals of existing weight loss medications known as GLP-1s, including injection pens that look like Novo Nordisk’s blockbuster diabetes drug Ozempic.
The ad claims those drugs are “priced for profits, not patients,” and points to Hims & Hers’ weight loss medications as “affordable” and “doctor-trusted” alternatives.
“We are complying with existing law and are happy to continue working with Congress and the new Administration to fix the broken health system and ensure that patients have choices for quality, safe, and affordable healthcare,” a Hims & Hers spokesperson told CNBC in a statement.
The senators do not mention Hims & Hers by name in their letter, but they do reference some of the visuals in the ad, including “imagery of an injection pen with distinctive characteristics reflective of an existing brand-name medication.”
“Nowhere in this promotion is there any side effect disclosure, risk, or safety information as would be typically required in a pharmaceutical advertisement,” the senators wrote. “Further, for only three seconds during the minute-long commercial does the screen flash in small, barely legible font, that these products are not FDA-approved.”
Hims & Hers began offering compounded semaglutide through its platform in May after launching a new weight loss program in late 2023. Semaglutide is the active ingredient in Ozempic and Wegovy, which can each cost around $1,000 a month without insurance.
Shares of Hims & Hers jumped over 170% last year, thanks to soaring demand for GLP-1s. They rose another 8% on Friday, lifting the company’s market cap to about $9.5 billion.
Compounded GLP-1s are typically much cheaper and can serve as an alternative for patients that are navigating complex supply hurdles and spotty insurance coverage. Hims & Hers sells compounded semaglutide for under $200 a month.
The FDA doesn’t review the safety and efficacy of compounded products, which are custom-made alternatives to brand drugs designed to meet a specific patient’s needs. Compounded products can also be produced when brand-name treatments are in shortage.
Semaglutide is currently in shortage, according to the FDA.
Sens. Durbin and Marshall said that advertisements for brand-name GLP-1 medications include “significant risk disclosures to patients about side effects and contraindications, including warnings about potential gallbladder, pancreas, vomiting, diarrhea, and other implications.”
A release on Durbin’s website says that the ad in question appears to exploit a loophole “regarding promotions of compounded drugs by telehealth companies.”
The senators said they believe the FDA may have the authority to take enforcement actions against marketing that could mislead patients, and they plan to introduce new legislation to address regulatory loopholes.
Employees package and sort express parcels at an e-commerce company on Nov. 1, 2024, around the Double 11 Shopping Festival in Lianyungang, Jiangsu Province of China.
Vcg | Visual China Group | Getty Images
President Donald Trump signed an executive order on Friday that puts a pause on his closing of the de minimis trade exemption, a provision commonly used by Chinese e-commerce companies Temu and Shein.
The order states that de minimis will be restored for small packages shipped from China, “but shall cease to be available for such articles upon notification by the Secretary of Commerce to the President that adequate systems are in place to fully and expediently process and collect tariff revenue” on those items.
Trump on Saturday suspended the exemption as part of new tariffs that include an additional 10% tax on Chinese goods. The nearly century-old exception, known as de minimis, has been used by many e-commerce companies to send goods worth less than $800 into the U.S. duty-free, creating a competitive advantage.
It was predicted that its removal could overwhelm U.S. Customs and Border Protection employees, as the mountain of low-value shipments already making their way into the U.S. would suddenly require formal processing.
De minimis has helped fuel an explosion in cheap goods being shipped from China into the U.S. CBP has said it processed more than 1.3 billion de minimis shipments in 2024. A 2023 report from the House Select Committee on the Chinese Communist Party found that Temu and Shein are “likely responsible” for more than 30% of de minimis shipments into the U.S., and “likely nearly half” of all de minimis shipments originate from China.
Critics of the de minimis provision say it’s provided an unfair advantage to Chinese e-commerce companies, and created an influx of packages that are “subject to minimal documentation and inspection,” raising concerns around counterfeit and unsafe goods.
The Biden administration proposed a new rule last September to curb the “overuse and abuse” of de minimis. The rule proposes to strengthen the CBP’s information collection requirements for de minimis shipments.