Connect with us

Business

Meta now has an AI chatbot. Experts say get ready for more AI-powered social media

Published

on

Meta now has an AI chatbot. Experts say get ready for more AI-powered social media

When you use Facebook Messenger these days, a new prompt greets you with this come-on: “Ask Meta AI anything.”

You may have opened the app to send a text to a pal, but Meta’s new artificial-intelligence-powered chatbot is tempting you with encyclopedic knowledge that‘s just a few keystrokes away.

Meta, the parent company of Facebook, has planted its home-grown chatbot on its Whatsapp and Instagram services. Now, billions of internet users can open one of these free social media platforms and draw on Meta AI’s services as a dictionary, guidebook, counselor or illustrator, among many other tasks it can perform — although not always reliably or infalliably.

“Our goal is to build the world’s leading AI and make it available to everyone,” said Mark Zuckerberg, the chief executive officer at Meta, as he announced the chatbot’s launch two weeks ago. “We believe that meta AI is now the most intelligent AI assistant that you can freely use.”

As Meta’s moves suggest, generative AI is making its way into social media. TikTok has an engineering team focused on developing large language models that can recognize and generate text, and they’re hiring writers and reporters who can annotate and improve the performance of these AI models. On Instagram’s help page it states, “Meta may use [user] messages to train the AI model, helping make the AIs better.”

Advertisement

TikTok and Meta did not respond to a request for comment, but AI experts said social media users can expect to see more of this technology influencing their experience — for better or possibly worse.

Part of the reason social media apps are investing in AI is that they want to become “stickier” for consumers, said Ethan Mollick, professor at the Wharton School of the University of Pennsylvania who teaches entrepreneurship and innovation. Apps like Instagram try to keep users on their platforms for as long as possible because captive attention generates ad revenue, he said.

At Meta’s first-quarter earnings call, Zuckerberg said it would take some time for the company to turn a profit from its investments in the chatbot and other uses of AI, but it has already seen the technology influencing user experiences across its platforms.

“Right now, about 30% of the posts on Facebook feed are delivered by our AI recommendation system,” Zuckerberg said, referring to the behind-the-scenes technology that shapes what Facebook users see. “And for the first time ever, more than 50% of the content people see on Instagram is now AI recommended.”

In the future AI won’t just personalize user experiences, said Jaime Sevilla, who directs Epoch, a research institute that studies AI technology trends. In fall 2022, millions of users were enraptured by Lensa’s AI capabilities as it generated whimsical portraits from selfies. Expect to see more of this, Sevilla said.

Advertisement

“I think you’re gonna end up seeing entirely AI-generated people who post AI-generated music and stuff,” he said. “We might live in a world where the part that humans play in social media is a small part of the whole thing.”

Mollick, author of the book “Co-intelligence: Living and Working with AI,” said these chatbots are already producing some of what people read online. “AI is increasingly driving lots of communication online,” he said. “[But] we don’t actually know how much AI writing is out there.”

Sevilla said generative AI probably won’t supplant the digital town square created by social media. People crave the authenticity of their interactions with friends and family online, he said, and social media companies need to preserve a balance between that and AI-generated content and targeted advertising.

Although AI can help consumers find more useful products in the daily lives, there’s also a dark side to the technology’s allure that can teeter into coercion, Sevilla said.

“The systems are gonna be pretty good at persuasion,” he said. A study just published by AI researchers at the Swiss Federal Institute of Technology Lausanne found that GPT-4 was 81.7% more effective than a human at convincing someone in a debate to agree. While the study has yet to be peer reviewed, Sevilla said that the findings were worrisome.

Advertisement

“That is concerning that [AI] might like significantly expand the capacity of scammers to engage with many victims and to perpetrate more and more fraud,” he added.

Sevilla said policymakers should be aware of AI’s dangers in spreading misinformation as the United States heads into another politically charged voting season this fall. Other experts warn that it’s not if, but how AI might play a role in influencing democratic systems across the world.

Bindu Reddy, CEO and co-founder of Abacus.AI, said the solution is a little more nuanced than banning AI on our social media platforms — bad actors were spreading hate and misinformation online well before AI entered the equation. For example, human rights advocates criticized Facebook in 2017 for failing to filter out online hate speech that fueled the Rohingya genocide in Myanmar.

In Reddy’s experience, AI has been good at detecting things such as bias and pornography on online platforms. She’s been using AI for content moderation since 2016, when she released an anonymous social network app called Candid that relied on natural language processing to detect misinformation.

Regulators should prohibit people from using AI to create deepfakes of real people, Reddy said. But she’s critical of laws like the European Union’s sweeping restrictions on the development of AI. In her view it’s dangerous for the U.S. to be caught behind competing countries, such as China and Saudi Arabia, that are pouring billions of dollars into developing AI technology.

Advertisement

So far the Biden administration has published a “Blueprint for an AI Bill of Rights” that offers suggestions for the safeguards that the public should have, including protections for data privacy and against algorithmic discrimination. It isn’t enforceable, though it hints at legislation that may come.

Sevilla acknowledged that AI moderators can be trained to have a company’s biases, leading to some views being censored. But human moderators have shown political biases too.

For example, in 2021 The Times reported on complaints that pro-Palestinian content was made hard to find across Facebook and Instagram. And conservative critics accused Twitter of political bias in 2020 because it blocked links to a New York Post story about the contents of Hunter Biden’s laptop.

“We can actually study like what kind of biases [AI] reflects,” Sevilla said.

Still, he said, AI could become so effective that it could powerfully oppress free speech.

Advertisement

“What happens when all that is in your timeline conforms perfectly to the company guidelines?” Sevilla said. “Is that the kind of social media you want to be to be consuming?”

Business

What Trump Gained, and Didn’t, From China

Published

on

What Trump Gained, and Didn’t, From China

Andrew here. With President Trump set to arrive back in Washington on Friday, we’re taking a hard look at what his high-stakes summit in Beijing actually achieved. The TL;DR: It didn’t lead to the “grand bargain” many had anticipated.

While there were optics of cooperation between Trump and Xi Jinping, concrete deals — including on Nvidia chips or tariffs — were few. Trump just said that he rejected a proposal from Xi, China’s leader, to help broker a peace between the U.S. and Iran, leaving the critical Strait of Hormuz effectively shut.

Ultimately, the president is coming home to rising oil prices and a slumping bond market.

President Trump departed Beijing a few hours ago, hailing “fantastic trade deals” struck during his two-day summit.

Still, many analysts and investors appear underwhelmed by a lack of details or breakthroughs on key issues like tariffs, Iran and tech restrictions. The summit seems to have fallen short of already diminished expectations.

Advertisement

For the 17 business leaders who accompanied Trump on the trip, the deal flow also appeared thinner than what was announced on his last presidential trip to China, in 2017.

Here are the highlights so far, Grady McGregor writes.

Nvidia and Citi apparently scored wins. Shares in Nvidia, the chipmaker, hit a record on Thursday on reports that Washington had cleared 10 Chinese companies to buy its H200 semiconductors.

That said, Beijing, which is looking to champion domestic rivals like Huawei, has not signaled it would be open to permitting the sales — an issue echoed on Friday by Jamieson Greer, the U.S. trade representative.

And on the eve of the summit, Beijing approved Citi’s application to operate a securities business in China, ending a yearslong regulatory application process. It is unclear whether the presence of Jane Fraser, the bank’s C.E.O., on the trip played any role in Beijing’s decision. Citi shares gained on Thursday.

Advertisement

Boeing landed an order for 200 aircraft, a deal Trump highlighted in a Fox News interview last night.

But shares in the plane maker fell sharply in premarket trading on Friday: The number was short of analysts’ forecasts of at least 300 planes.

The Board of Trade looks like a go. The Washington-Beijing body would manage trade in sectors such as aviation, energy, medical equipment and agriculture. Greer said it would aim to reduce tariffs on roughly $30 billion worth of goods.

He added that he expected the tariff truce the countries struck last fall in South Korea to be extended.

What’s still unclear:

Advertisement

Major cryptocurrency regulation clears a key hurdle. The Senate Banking Committee passed the Clarity Act, which has been promoted by crypto companies and investors like the venture capital firm Andreessen Horowitz. The bill heads to the full Senate, where it faces a less certain fate.

Federal prosecutors will drop criminal charges against India’s richest man. The move to end the case against the businessman Gautam Adani came after one of his lawyers — Robert Giuffra, who is also one of President Trump’s personal lawyers — met with Justice Department officials, The Times reports. (A presentation by Giuffra said that Adani was willing to invest $10 billion in the U.S., though sources told The Times that the withdrawal of charges wasn’t tied to the offer.) A settlement in a parallel case by the S.E.C. was announced Thursday in which Adani agreed to pay $6 million.

Bill Ackman bets big on Microsoft. The billionaire financier said on Friday that he had acquired a major stake in the tech giant and that he believed in the long-term prospects of its productivity software and its spending on A.I. Other hedge fund managers have bet the opposite: TCI, the firm run by Chris Hohn, recently sold off an $8 billion stake in Microsoft.

The high-stakes legal showdown between Elon Musk and OpenAI is finally headed to the nine-person jury.

Over more than seven hours of closing arguments, lawyers for each side sought to paint the other as untrustworthy.

Advertisement

Here are some of the highlights of Thursday’s proceedings.

Can anyone trust Sam Altman? That was again the central attack by Steven Molo, Musk’s lead lawyer, who has argued that Altman, the OpenAI chief, deceived Musk, a fellow founder, about plans to convert the company from nonprofit to for-profit.

Molo told jurors that five witnesses had called Altman a “liar,” and he hammered home his point with a creative metaphor:

Imagine that you’re on a hike, and you come upon one of those wooden bridges that you see on a trail, and it’s over a gorge. There’s a river that’s 100 feet below and it looks a little scary, but a woman standing by the entry to the bridge says, “Don’t worry, the bridge is built on Sam Altman’s version of the truth.” Would you walk across that bridge? I don’t think many people would.

Can jurors trust Musk’s version of events? OpenAI’s lawyers, from the law firm Wachtell, Lipton, Rosen & Katz, argued that the billionaire knew about the company’s plans for for-profit conversion earlier than he admitted to and that the statute of limitations for his claims had passed.

Referring to Musk’s claim that he hadn’t read most of a 2018 email about OpenAI’s plans to seek outside investment, Sarah Eddy, a lawyer for OpenAI, said:

Advertisement

Here you have one of the most sophisticated businessmen in the history of the world and he claims he didn’t read a four-page summary term sheet.

The outcome of the trial could drastically alter the A.I. landscape. If OpenAI loses, its operations could be disrupted at a time when rivals are gaining steam.

The artificial intelligence boom has been a tale of haves and have-nots. Some companies have benefited mightily, most recently the chip maker Cerebras, whose stock shot up 68 percent in its debut. But many enterprise software providers have been walloped.

One of them was Figma, the design-software maker whose shares have tumbled since it went public last year. But as it reported strong quarterly earnings on Thursday, its C.E.O., Dylan Field, spoke with Michael de la Merced about why he believed his company was poised to survive, and even thrive. Here are our takeaways after the conversation.

Remember the “SaaSpocalypse”? Referring to “software-as-a-service,” it referred to investors’ worries that tools like Anthropic’s Claude Code would devastate the entire category of subscription-based software companies, like Figma.

Figma appears to have dispelled at least some of those worries:

Advertisement

The company’s results held up after an A.I.-related change in pricing. For most of its existence, Figma charged companies per user (known as seat-based pricing). But A.I. agents that can do work once reserved for humans promise to drastically reduce how many “seats” customers need to pay for.

In mid-March, Figma switched to a system in which it charged users for how much A.I. they used past a certain amount. The company said that more than 75 percent of its business users kept using A.I. tools despite the cap.

The result: Shares in Figma are up more than 10 percent in premarket trading since the report.

“Market narratives are market narratives,” Field said to DealBook about the SaaSpocalypse sell-off, playing down the investor concern while pointing out Figma’s strong performance.

“The way we see it, A.I. is going to create more software than ever,” he said. He added, “Design matters.”

Advertisement

But Field remains on guard. Makers of A.I. models have muscled into Figma’s territory, notably Anthropic, which in March introduced Claude Design, a tool seen as a competitor of sorts. (Only three days before, Mike Krieger, a senior Anthropic executive, resigned from Figma’s board; Field reportedly complained about the situation.)

“You have to take a company like Anthropic seriously,” Field told DealBook.

The musical playlist for Thursday’s state dinner in Beijing for President Trump drew big buzz on social media. It contained some Trump favorites, including the Village People hit “Y.M.C.A.”


Every week, we’re asking a leader how he or she uses artificial intelligence. This week, Jeremy Allaire, who leads the stablecoin issuer Circle, told Sarah Kessler that he had built a “C.E.O. prioritizer.” The interview has been condensed and edited for clarity.

How do you personally use A.I. at home or work?

Advertisement

One interesting one is a C.E.O. prioritizer. If there’s a request for me to meet someone or do something, you go to the agent and it interrogates you about it and does background research. Then it assigns a one-to-five score, with one being “Completely ignore it” and five being “This is a highly strategic use of your time.”

Circle wants to be part of the infrastructure that helps A.I. agents spend money. Tell me more about that.

The primary units of work in the economic system are going to be executed by A.I. agents. And increasingly, it’s going to be agents that are operating in teams.

You need an economic system to support that. We need a way for one agent to access and use the services of another agent. For example, you might have research data in a particular domain of biology, and I want to make that available to A.I.s to consume. And it’s going to be 5 cents, 10 cents. Whatever it is, you receive that payment, and the A.I. then can consume that data and use it.

And this transaction would take place via stablecoin and not dollars, because there is less friction and these are tiny transactions?

Advertisement

There’s no payment system in the world except for something like USDC that can conduct a transaction for a fraction of a penny. Or even 5 cents or 10 cents. And it’s all programmable.

You said on your latest earnings call that 85 percent of your employees are using A.I. coding and automation tools. What does that look like?

We’re able to basically go through the entire software life cycle with A.I. agents conducting work. Agents are seeing feature requests, picking them up, coding and submitting the code for review. We have other agents that perform code review. Humans then obviously come in to do subsequent reviews.

What about outside of engineering?

It’s in every single function. If you want to build a creative strategy for a campaign, there’s a whole agentic workflow. If you are creating public communications content — we’re a regulated company, so we have very strict guidelines — there’s an A.I. that will vet all of your content and point out the issues with it.

Advertisement

Deals

  • Investors led by Egon Durban, a C.E.O. of the tech investment firm Silver Lake, have reportedly struck a deal to buy 25 percent of the Las Vegas Raiders at a $9.9 billion valuation. (CNBC)

  • Michael Carr, a longtime top M.&A. banker at Goldman Sachs, died on Tuesday. He was 68. (Bloomberg)

Politics, policy and regulation

Best of the rest

  • Boeing and Toyota are said to have donated $1 million each to fund a reality-TV video series starring the transportation secretary, Sean Duffy. (WSJ)

  • “In a City of Big Dreams, Many Young Adults See a Cloudy Future” (NYT)

We’d like your feedback! Please email thoughts and suggestions to dealbook@nytimes.com.

Advertisement
Continue Reading

Business

Why the infuriatingly catchy Kars4Kids jingle got yanked off the air in California

Published

on

Why the infuriatingly catchy Kars4Kids jingle got yanked off the air in California

The frustratingly unforgettable Kars4kids jingle, which has been worming its way into listeners’ brains for decades, is officially banned from the airwaves in California.

While the 1-877-KARS4KIDS song has been called one of the most memorable jingles in history, a court has ruled it is misleading.

A California man took the group behind it to court, saying he donated an old car to Kars4Kids, thinking its value would be used to help underprivileged children. He didn’t know the money generated was used to support Oorah, a Jewish organization that helps fund young adult trips to Israel.

An Orange County court judge ruled late last week that the New Jersey-based group’s advertisements were misleading because they omitted the company’s religious affiliation and hid the charity’s true mission.

Advertisement

The charity organization violated state laws against false advertising and unfair competition, the court ruled.

“The failure to disclose that funds benefit adults and families — and that this support is contingent upon a specific religious affiliation — is a material omission,” the ruling states.

Kars4Kids must pull its advertisements from the state within 30 days. Any new advertisement in California must clearly disclose the nonprofit’s religious affiliation and specify for whom the money will be used, the court ruled.

A Kars4Kids spokesperson said the ruling is deeply flawed, and the organization will appeal.

“We believe this case was nothing more than a lawyer-driven attempt to siphon off charitable funds for their own gain,” the spokesperson said. “The law and the facts are clearly on our side.”

Advertisement

The jingle first aired in the 1990s and has been loved and loathed by listeners ever since.

It has been the subject of talk show commentary and featured in “The Simpsons.”

Most donations go to help Jewish youth and families, the company’s chief operating executive, Esti Landau, said during her testimony.

Oorah runs a matchmaking program for Jewish youth and funds gap year trips to Israel for 17- and 18-year-olds. The company also used donations to purchase a $16.5-million building in Israel.

“The evidence also shows that children, especially needy or underprivileged children, are not the recipients of the proceeds of the donations,” the ruling states.

Advertisement

Kars4Kids has made it easier to donate old cars to benefit children and families across the country, which includes continued support throughout young adulthood, the company spokesperson said.

This isn’t the first time Kars4Kids has faced accusations of misleading listeners. Oregon, Pennsylvania and other states have also found the charity organization has misleading solicitation practices.

Californians account for a quarter of the company’s funds, yet the nonprofit has limited programs in the state, according to court documents. The organization claims to help thousands of children, including hundreds in California, according to a Kars4Kids spokesperson.

The charity’s infamous tune was catchy enough to convince California resident Bruce Puterbaugh to donate a 2001 Volvo XC. The car was nonoperational and not under his name, but was left in his care.

The car was valued at $250, and Puterbaugh said he felt deceived when he found out the money wouldn’t help young children. He originally sued the company in 2021.

Advertisement

“I feel taken advantage of by the ad and information that was not there,” Puterbaugh said in his testimony.

A donor would have to navigate the nonprofit’s website to learn about its religious mission.

“These omissions are inherently deceptive,” the court ruling states. “Broadcasting this jingle repeatedly over two decades is fraudulent.”

A Kars4Kids spokesperson said that the company’s website clearly states its Jewish affiliation.

The court sided with Puterbaugh and ordered the nonprofit to pay him $250 as restitution for his donated vehicle.

Advertisement
Continue Reading

Business

Some Medicare Patients Can Now Get Free CBD

Published

on

Some Medicare Patients Can Now Get Free CBD

The Trump administration has been making headlines for taking steps to loosen restrictions around cannabis, including legalizing it for medical use. Now it is beginning an experiment that places cannabis even more squarely into mainstream health care: thousands of Medicare patients soon will be able to get CBD, a nonintoxicating component, for free.

“ONE in FIVE adults used it in the past year, and many say it improved their chronic pain enormously,” President Trump wrote on social media last month in a post cheering the program.

The aim is to gather real-life evidence showing whether CBD can improve patients’ quality of life and, by extension, reduce health care costs, administration officials say.

CBD products are already popular with some Medicare-age patients. A 2024 study in Clinical Gerontologist found that 14.3 percent of patients 65 and older had used them in the past year. Patients usually purchase over-the-counter gummies and tinctures to ease anxiety, insomnia and chemo-related nausea.

“Millions of older adults are already integrating cannabinoid products into their health care routines, yet the health care system has almost no infrastructure to understand what they are spending, why they are using these products, or whether these expenditures reduce other health care costs,” said Sasha Kalcheff-Korn, the executive director of Realm of Caring, a nonprofit group that conducts research and promotes cannabinoid therapies.

Advertisement

Despite Mr. Trump’s ebullient endorsement, many doctors worry about encouraging the use of unapproved supplements to geriatric patients, who typically have multiple medical conditions and already take many medications, some of which could interact with CBD products to detrimental effect. Still, their concerns would be eased somewhat, they say, if patients collaborated with doctors on appropriate dosing, which is another goal of the government initiative.

”I believe that CBD should be available to all seniors as part of their health care, recommended by a provider with knowledge of cannabinoid medicine,” said Dr. Melanie Bone, the director of medical cannabinoid therapies at MorseLife, a senior residence in West Palm Beach, Fla. “It may help with a number of ailments of aging, and has almost no downside. But CBD is not a panacea. The only way to know if it works is to try.”

CBD, or cannabidiol, one of the most prominent compounds in the cannabis sativa plant, is nonintoxicating and known for its soothing effects on the central nervous system. Many CBD products are made from hemp, a legal strain of cannabis that is rich in CBD and has only small amounts of the intoxicating compound, THC. The Medicare program restricts the amount of THC that can be in hemp-derived CBD to 3 milligrams per serving.

In recent years, CBD has become increasingly attractive to older patients. Results from studies are mixed to positive. But many of the doses evaluated contained more THC than those allowed by the Medicare guidelines. Most researchers have noted the need for more rigorous gold-standard trials.

Mr. Trump’s assertion that one in five adults use CBD products, many for chronic pain, which was also included in supporting documents for an executive order announcing the program, appears to conflate self-reported surveys and polls that broadly address adult use of medical cannabis or CBD.

Advertisement

But a chief benefit of CBD that some studies do underscore is that many seniors use the products to replace opioids for pain and benzodiazepines for anxiety and insomnia, which can have troubling side effects.

The new Medicare program mandates that the CBD be given to patients only by doctors, who regularly review their medical history and reactions to the products.

One of the main goals is to learn whether CBD can help older people feel better enough to get off, or avoid starting, prescriptions for pain, nausea, sleep and anxiety. The hope is that CBD could help prevent more expensive medical interventions that those drugs can lead to. Opioids, for example, can prompt dizziness, constipation, overdoses and trips to the emergency department.

Only a small subset of Medicare recipients — those who participate in a type of health care network called an Accountable Care Organization — will initially be eligible for the benefit. So far, just five large groups have been approved to offer CBD. By January, 2027, CBD will be offered to patients in all 74 ACO groups.

The participating organizations have providers across an array of states, including Oklahoma, Texas, Wisconsin, West Virginia and Arizona. Currently, only patients affiliated with programs in New York and Florida patients have begun receiving CBD products, according to a spokesperson for the Centers for Medicare and Medicaid Services.

Advertisement

Those doctors must buy the CBD products up front, spending up to $500 per patient a year. They must agree to screen patients and products carefully, and collect real-time data on how the CBD affects patients.

They will not be directly reimbursed for the CBD. In the incentive-based structure, these groups receive a budget from Medicare. Those that come in under budget by improving patient quality of life and reducing costs, now additionally equipped with CBD as a tool, will receive a percentage of those savings.

Yes.

Late last year, Congress passed a measure that could remove from the U.S. market most CBD products, including those that doctors suggest for patients.

That is because many CBD products contain far more THC and other synthetic, intoxicating compounds than Congress intended in 2018, when it created the legal definition of hemp, to distinguish it from marijuana. Many of those amped-up CBD items, packaged to look like candy, have led to calls to poison centers.

Advertisement

In reaction, Congress placed severe limits on hemp last year that are set to take place in November. Under those restrictions, Cornbread Hemp, a Kentucky-based company with a contract to supply CBD for the new program, will not be able to do so, because its products’ THC content is above the new limits. A patchwork of bills introduced in the Senate and the House are trying to slow or rewrite what amounts to a looming hemp ban.

In his social media post last month, Mr. Trump urged Congress to act.

“Please get it done, and SOON,” he wrote.

Continue Reading
Advertisement

Trending