Soon after the new administration arrived, things began to go missing from the White House website.

They weren’t just the partisan policy platforms that typically disappear during a presidential transition. Informational pages about the Constitution and past presidents, up in various forms since President George W. Bush was in office, all vanished.

Thousands of other government web pages had also been taken down or modified, including content about vaccines, hate crimes, low-income children, opioid addiction and veterans, before a court order temporarily blocked part of the sweeping erasure. A Justice Department database tracking criminal charges and convictions linked to the Jan. 6, 2021, attack on the Capitol was removed. Segments of data sets are gone, some of the experts who produced them were dismissed, and many mentions of words like “Black,” “women” and “discrimination” have evaporated.

President Trump’s team is selectively stripping away the public record, reconstructing his preferred vision of America in the negative space of purged history, archivists and historians said. As data and resources are deleted or altered, something foundational is also at risk: Americans’ ability to access and evaluate their past, and with it, their already shaky trust in facts.

“This is not a cost-cutting mechanism,” said Kenny Evans, who studies science and technology policy at Rice University’s Baker Institute for Public Policy and runs the White House Scientists Archive at the school. “This slide toward secrecy and lack of transparency is an erosion of democratic norms.”

Anna Kelly, a White House spokeswoman, said on X that the disposal process was standard practice for old courtesy copies of paperwork that were largely backed up on classified computer systems. In an emailed statement, she did not address concerns about the removed records, but said that the president regularly communicated with news outlets and directly with the public and was “leading the most transparent administration in history.”

“He is adding transparency by exposing the vast waste, fraud and abuse across the federal government and restoring accountability to taxpayers,” she said.

The campaign of deletion does more than amplify the administration’s policy priorities — it buries evidence of the alternatives in a MAGA-branded memory hole. Several information experts said that Mr. Trump’s executive orders have authoritarian overtones, reminiscent of when Russia cloned Wikipedia, an online encyclopedia, and stripped it of unflattering material. Information experts and civil rights groups fear that a historical vacuum could jeopardize accountability and breed mistrust, especially in an already hostile political environment for researchers who are trying to fight disinformation.

“There are tectonic plates that are shifting, and it’s a new version of truth that is being portrayed, and that, I think, is the most profound danger we have ever faced as a country,” said Laurence H. Tribe, a constitutional scholar and professor emeritus at Harvard Law School.

Even Utah’s Republican lieutenant governor called on Mr. Trump to “bring back our history” after the first American woman to legally vote was removed from the website for Arlington National Cemetery, along with a section on other notable women (her profile is once again available, but the women’s history section is not). References to transgender people disappeared from the National Park Service’s web page for the Stonewall National Monument.

Mr. Trump is not known as an enthusiast of document preservation: Past employees have described his penchant for ripping up documents and flushing papers down the toilet.

But his administration has surfaced some government data. In March, the National Archives released some 64,000 documents about the assassination of President John F. Kennedy, including accounting records that contained the Social Security numbers of dozens of government workers from the late 1970s, some of them still living.

The restructuring effort led by Elon Musk through his Department of Government Efficiency, which had been caught in a series of high-profile errors, tried to delete or obscure the mistakes before reversing course last month and adding more details that fact-checkers could use to confirm its claims about the savings it had achieved from canceling federal grants.

The historical record, however, remains under intense pressure and not just from the government.

Mr. Musk has a vendetta against Wikipedia, which the billionaire derided as “Wokepedia” last year. He called the encyclopedia, which is written and edited by volunteers from the general public, “an extension of legacy media propaganda” after an entry described a gesture he had made during Mr. Trump’s recent inauguration as being “compared to a Nazi salute.” Jimmy Wales, a founder of Wikipedia, pushed back on the social media platform X, saying “that’s fact. Every element of it.”

Data Foundation, a think tank, said in a report last month that changes in federal evidence gathering are coinciding with similar shifts in the private data sector. Those include more than 2,000 layoffs and other departures in March and several analysis firms shutting down entirely. A year ago, Google also removed links to cached pages from its search results, stripping away a longtime feature that helped researchers and others track changes on websites.

Resources from the government have become especially important as researchers find themselves limited or cut off from data reserves kept by social media companies, said Samuel Woolley, the disinformation studies chair at the University of Pittsburgh.

“The idea that suddenly we no longer need oversight or access to the information that allows us to conduct oversight is worrying,” he said. “Getting rid of public records and people who study things like influence operations amounts to a kind of censorship by omission.”

Outside the government, many archivists are now rushing to preserve endangered material.

The Data Rescue Project, which launched in February, is cataloging preservation efforts and backed up government data sets. Since 2008, the End of Term Web Archive has conducted “a comprehensive harvest” of federal government domains and chronicled changes from administration to administration. Initiatives like the Environmental Data & Governance Initiative and the Open Environmental Data Project are storing copies of government climate data.

Another key participant: The Internet Archive, a 28-year-old nonprofit library housed in a stately former Christian Science church in San Francisco. Some 140 workers, mostly engineers, archive more than a billion URLs a day with help from partners such as Cloudflare, WordPress, Reddit and Wikipedia’s parent organization, Wikimedia. The work is funded through donations and web archiving agreements with more than 1,300 schools, museums and libraries.

The Archive has collected more than 700,000 terabytes of archived web pages as one of the partners working on the End of Term project, identifying more than 150,000 government pages that have gone offline since the inauguration.

“What we’re seeing this time around is unprecedented, both in terms of the scope and the scale of the web-based resources that are being taken offline, and material on those pages that is being changed,” said Mark Graham, the director of the Wayback Machine, a digital repository operated through the Internet Archive.

The Archive has faced difficulties in recent years, such as copyright lawsuits from record labels and book publishers seeking hundreds of millions of dollars in damages (the organization had a $28 million budget in 2023). It has also been targeted by cyberattacks.

The Trump administration, however, has not been an obstacle. Mr. Musk has called the archive “awesome” and “a public good that should exist,” even as he complained about “a ton of negative” content that concerned him.

In February, government lawyers argued that the removal of information from the C.D.C. website caused limited harm because the scrubbed pages could still be viewed on the Wayback Machine. A federal judge disagreed, noting that the site does not capture every page, and the ones that are archived do not appear on search engines and can only be found using their original URL.

Mr. Graham, an Air Force veteran who can rattle off URLs from memory, said he has worked seven days a week with few breaks since Mr. Trump took office.

“We’ve seen examples throughout history and all over the world where governments attempt to change culture, change the values of a population by changing and/or restricting access to information,” he said. “I think we still see that to this day.”

When President Trump announced sweeping tariffs this week, some of the biggest tech companies had obvious reasons for worry.

Apple, Dell and Oracle — which rely on hardware and global supply chains that are in the direct line of fire from tariffs — saw their shares go into free-fall. But there was another big tech company whose stock took a pummeling even though its core business has little to do with hardware: Meta.

Shares of the company, which owns Facebook, Instagram and WhatsApp, fell $52 to $531.62 on Thursday and were down again on Friday. In total, Meta shed a whopping 9 percent of its market capitalization on Thursday.

The reasons for Meta’s slide are less obvious. But close watchers of the social networking and metaverse company know it is just as vulnerable to Mr. Trump’s trade actions as some of its Silicon Valley peers, even if the details are more complicated. Here’s why.

What gives with Meta? It isn’t in the same hardware businesses as Apple or Nvidia.

That’s not entirely true, but for our purposes let’s look at Meta’s main business: digital advertising.

Meta rakes in billions of dollars in revenue by selling ads across Facebook and Instagram. Some of those advertisers are large brands, including Procter & Gamble, L’Oreal, McDonald’s and Nestlé. Those companies buy ads on Facebook for so-called brand awareness campaigns. Think of it as a way of nudging people to buy a specific product like Q-Tips instead of generic cotton swabs the next time they go to the store.

But a vast majority of Meta’s advertisers are small and medium-size businesses.

Those companies buy a different kind of ad, called “direct response advertising.” These ads typically encourage an action of some sort, like downloading a company’s app or buying a kitchen gadget featured on an Instagram video.

E-commerce transactions like these make up an enormous amount of Meta’s very successful online advertising business. Susan Li, Meta’s chief financial officer, said in an earnings call this year that online commerce ads were the “largest contributor to year-over-year growth” to the company’s advertising revenue.

So what does that mean?

The effect of tariffs on Meta’s ad business is simple. Many of its small and medium-size advertisers are from all across the world. President Trump’s tariffs will instantly make it more expensive for them to sell their products to customers in the United States.

That is likely to lead to a pullback in overall purchases from consumers and fewer people buying products from Facebook and Instagram. That could, in turn, lead brands to spend less on advertising across those apps.

This seems somewhat hypothetical. Could that really lead to such a big stock drop for Meta?

Meta has additional complicating factors that may affect its business more than other advertising companies.

Last year, the company disclosed that 10 percent of its revenue in 2023 was from Chinese companies spending heavily on advertising across Facebook and Instagram, an ad blitz aimed at garnering a foothold in lucrative Western markets.

Much of that growth was fueled by the explosive expansion of the fast-fashion company Shein — which is based in Singapore but has a supply chain that is largely in China — and the e-commerce app Temu, a low-cost, Amazon-like company owned by the Chinese e-commerce conglomerate Pinduoduo. Temu was estimated to have spent $3 billion in marketing costs in 2023 alone, according to estimates from Bernstein Research.

Chinese companies and goods have been hit hard by President Trump’s tariffs. In addition, Mr. Trump eliminated the “de minimis exemption,” which had exempted exporters sending goods valued at or less than $800 from having to pay duties. The exemption was essential to the Temu and Shein business model of selling low-cost goods to Americans.

If Mr. Trump’s tariffs stick, they could drastically hurt these exporters of cheap Chinese goods, which means they could slash their advertising on Facebook and Instagram.

Just how exposed is Meta?

In an investor call last year, Ms. Li defended the company’s exposure to any fluctuation in spending by Temu and Shein.

She said two-thirds of Meta’s Chinese ad revenue came from advertisers “outside the top 10 spenders in that country in 2023.” Her point being: Even if Temu and Shein pulled back, many other Chinese advertisers were still buying Facebook and Instagram ads.

Unfortunately for Meta, that broad base of advertisers is no hedge against Mr. Trump’s tariffs, which will affect all Chinese ad buyers.

“Because their Chinese ad revenue is so evenly distributed, it’s actually worse for them now,” said Eric Seufert, an independent mobile advertising analyst who follows Meta. “They don’t just have to worry about Temu or Shein dropping off. They have to worry about everyone.”

Meta did not respond to requests for comment.

Oof.

To be fair, Meta isn’t alone. E-commerce tech companies like Shopify and Stripe could face headwinds if global trade slows. Google and Amazon also have enormous ad businesses that could be hampered by a pullback in Chinese companies’ spending.

We will hear Meta’s defense soon enough. The company is expected to answer investors’ questions when it reports quarterly earnings this month.

The year is 2027. Powerful artificial intelligence systems are becoming smarter than humans, and are wreaking havoc on the global order. Chinese spies have stolen America’s A.I. secrets, and the White House is rushing to retaliate. Inside a leading A.I. lab, engineers are spooked to discover that their models are starting to deceive them, raising the possibility that they’ll go rogue.

These aren’t scenes from a sci-fi screenplay. They’re scenarios envisioned by a nonprofit in Berkeley, Calif., called the A.I. Futures Project, which has spent the past year trying to predict what the world will look like over the next few years, as increasingly powerful A.I. systems are developed.

The project is led by Daniel Kokotajlo, a former OpenAI researcher who left the company last year over his concerns that it was acting recklessly.

While at OpenAI, where he was on the governance team, Mr. Kokotajlo wrote detailed internal reports about how the race for artificial general intelligence, or A.G.I. — a fuzzy term for human-level machine intelligence — might unfold. After leaving, he teamed up with Eli Lifland, an A.I. researcher who had a track record of accurately forecasting world events. They got to work trying to predict A.I.’s next wave.

The result is “AI 2027,” a report and website released this week that describes, in a detailed fictional scenario, what could happen if A.I. systems surpass human-level intelligence — which the authors expect to happen in the next two to three years.

“We predict that A.I.s will continue to improve to the point where they’re fully autonomous agents that are better than humans at everything by the end of 2027 or so,” Mr. Kokotajlo said in a recent interview.

There’s no shortage of speculation about A.I. these days. San Francisco has been gripped by A.I. fervor, and the Bay Area’s tech scene has become a collection of warring tribes and splinter sects, each one convinced that it knows how the future will unfold.

Some A.I. predictions have taken the form of a manifesto, such as “Machines of Loving Grace,” a 14,000-word essay written last year by Dario Amodei, the chief executive of Anthropic, or “Situational Awareness,” a report by the former OpenAI researcher Leopold Aschenbrenner that was widely read in policy circles.

The people at the A.I. Futures Project designed theirs as a forecast scenario — essentially, a piece of rigorously researched science fiction that uses their best guesses about the future as plot points. The group spent nearly a year honing hundreds of predictions about A.I. Then, they brought in a writer — Scott Alexander, who writes the blog Astral Codex Ten — to help turn their forecast into a narrative.

“We took what we thought would happen and tried to make it engaging,” Mr. Lifland said.

Critics of this approach might argue that fictional A.I. stories are better at spooking people than educating them. And some A.I. experts will no doubt object to the group’s central claim that artificial intelligence will overtake human intelligence.

Ali Farhadi, the chief executive of the Allen Institute for Artificial Intelligence, an A.I. lab in Seattle, reviewed the “AI 2027” report and said he wasn’t impressed.

“I’m all for projections and forecasts, but this forecast doesn’t seem to be grounded in scientific evidence, or the reality of how things are evolving in A.I.,” he said.

There’s no question that some of the group’s views are extreme. (Mr. Kokotajlo, for example, told me last year that he believed there was a 70 percent chance that A.I. would destroy or catastrophically harm humanity.) And Mr. Kokotajlo and Mr. Lifland both have ties to Effective Altruism, another philosophical movement popular among tech workers that has been making dire warnings about A.I. for years.

But it’s also worth noting that some of Silicon Valley’s largest companies are planning for a world beyond A.G.I., and that many of the crazy-seeming predictions made about A.I. in the past — such as the view that machines would pass the Turing Test, a thought experiment that determines whether a machine can appear to communicate like a human — have come true.

In 2021, the year before ChatGPT launched, Mr. Kokotajlo wrote a blog post titled “What 2026 Looks Like,” outlining his view of how A.I. systems would progress. A number of his predictions proved prescient, and he became convinced that this kind of forecasting was valuable, and that he was good at it.

“It’s an elegant, convenient way to communicate your view to other people,” he said.

Last week, Mr. Kokotajlo and Mr. Lifland invited me to their office — a small room in a Berkeley co-working space called Constellation, where a number of A.I. safety organizations hang a shingle — to show me how they operate.

Mr. Kokotajlo, wearing a tan military-style jacket, grabbed a marker and wrote four abbreviations on a large whiteboard: SC > SAR > SIAR > ASI. Each one, he explained, represented a milestone in A.I. development.

First, he said, sometime in early 2027, if current trends hold, A.I. will be a superhuman coder. Then, by mid-2027, it will be a superhuman A.I. researcher — an autonomous agent that can oversee teams of A.I. coders and make new discoveries. Then, in late 2027 or early 2028, it will become a superintelligent A.I. researcher — a machine intelligence that knows more than we do about building advanced A.I., and can automate its own research and development, essentially building smarter versions of itself. From there, he said, it’s a short hop to artificial superintelligence, or A.S.I., at which point all bets are off.

If all of this sounds fantastical … well, it is. Nothing remotely like what Mr. Kokotajlo and Mr. Lifland are predicting is possible with today’s A.I. tools, which can barely order a burrito on DoorDash without getting stuck.

But they are confident that these blind spots will shrink quickly, as A.I. systems become good enough at coding to accelerate A.I. research and development.

Their report focuses on OpenBrain, a fictional A.I. company that builds a powerful A.I. system known as Agent-1. (They decided against singling out a particular A.I. company, instead creating a composite out of the leading American A.I. labs.)

As Agent-1 gets better at coding, it begins to automate much of the engineering work at OpenBrain, which allows the company to move faster and helps build Agent-2, an even more capable A.I. researcher. By late 2027, when the scenario ends, Agent-4 is making a year’s worth of A.I. research breakthroughs every week, and threatens to go rogue.

I asked Mr. Kokotajlo what he thought would happen after that. Did he think, for example, that life in the year 2030 would still be recognizable? Would the streets of Berkeley be filled with humanoid robots? People texting their A.I. girlfriends? Would any of us have jobs?

He gazed out the window, and admitted that he wasn’t sure. If the next few years went well and we kept A.I. under control, he said, he could envision a future where most people’s lives were still largely the same, but where nearby “special economic zones” filled with hyper-efficient robot factories would churn out everything we needed.

And if the next few years didn’t go well?

“Maybe the sky would be filled with pollution, and the people would be dead?” he said nonchalantly. “Something like that.”

One risk of dramatizing your A.I. predictions this way is that if you’re not careful, measured scenarios can veer into apocalyptic fantasies. Another is that, by trying to tell a dramatic story that captures people’s attention, you risk missing more boring outcomes, such as the scenario in which A.I. is generally well behaved and doesn’t cause much trouble for anyone.

Even though I agree with the authors of “AI 2027” that powerful A.I. systems are coming soon, I’m not convinced that superhuman A.I. coders will automatically pick up the other skills needed to bootstrap their way to general intelligence. And I’m wary of predictions that assume that A.I. progress will be smooth and exponential, with no major bottlenecks or roadblocks along the way.

But I think this kind of forecasting is worth doing, even if I disagree with some of the specific predictions. If powerful A.I. is really around the corner, we’re all going to need to start imagining some very strange futures.

Taiwan, the center of the global supply chain for computer chips, woke up on Thursday to the news that President Trump had put new 32 percent tariffs on the island’s exports to the United States. Except for semiconductors.

The decision not to impose tariffs on the chip sector does not mean they won’t be coming for Taiwan or anywhere else, including South Korea, another major source of chips.

Taiwanese companies have spent decades and billions of dollars building up a network of factories that conduct the complex process of etching tiny circuits onto pieces of silicon.

These chips — and a broad range of electronic devices that contain them — are Taiwan’s main exports. And they are increasingly the focal point of the Taiwanese-U.S. geopolitical relationship, which has undergone a markedly transactional shift since Mr. Trump took office.

Mr. Trump previously said Taiwan had gained an unfair dominance in making semiconductors and threatened to impose tariffs on the sector. He has also accused Taiwan, which depends on the United States for political support against China’s claims that Taiwan is part of its territory, of spending too little on its own security.

Officials and businesses in Taiwan have been scrambling to soften the blow of Washington’s tariff threats. Last month, President Lai Ching-te said Taiwan was interested in buying natural gas from a long-stalled project in Alaska.

Weeks earlier, Taiwan Semiconductor Manufacturing Company, the world’s largest chipmaker, said it would spend $100 billion in the United States to expand its operations in Arizona. TSMC announced plans for the plant during Mr. Trump’s first term and got a big financial assist under former President Joseph R. Biden Jr.

On Wednesday, in announcing the tariffs on Taiwan, Mr. Trump praised TSMC for investing in the United States. He and his aides are hoping other chip companies that committed to invest in U.S. operations during the Biden administration, like the South Korean giants Samsung and SK Hynix and Global Wafers of Taiwan, will pledge to spend even more.

Semiconductors are a complicated target for tariffs because the supply chain for making them is both global and extremely specialized. While most advanced chips are manufactured in Taiwan, many are then sent to another country, like Malaysia, for testing. Next, the chips might be put into iPhones or artificial intelligence servers in Mexico or China before those devices are sold to people all over the world.

“Very few semiconductors are actually directly imported into the United States; most are incorporated into a final product,” said Jimmy Goodrich, a senior adviser for technology analysis at the RAND Corporation.

Martin Chorzempa, a senior fellow at the Peterson Institute for International Economics, added: “It’s much more tricky than saying, ‘We are going to slap a tariff on steel.’”

Even the chips that TSMC makes at its Arizona factory have to leave the United States to be packaged into other devices before they make it into the hands of American consumers, said Ming-Yen Ho, a nonresident fellow at the Research Institute for Democracy, Society and Emerging Technology, a government-funded think tank in Taipei.

“Any chip that’s made in Arizona right now will have to leave the U.S. for some time before going back,” Mr. Ho said. “It’s just a fact of the global chip supply chain right now.”

Mr. Trump on Wednesday announced a 32 percent tariff on Taiwanese goods exported to the United States, where Taiwan directly sends nearly a quarter of its exports. Besides chips, which will not be taxed, Taiwan mainly exports electronic devices and components for them. The American Chamber of Commerce in Taiwan said Taiwan played an indispensable role in the U.S. economy and urged officials in Washington and Taipei to strengthen ties.

The Taiwanese government on Thursday condemned the tariffs as unreasonable and unfair to Taiwan. The government would lodge a strong protest with the Office of the United States Trade Representative, Lee Hui-chih, a spokeswoman for Taiwan’s cabinet, said in a statement.

Taiwan’s exports to the United States have increased in recent years, reflecting increased demand for Taiwanese electronics and advanced technology including semiconductors, the statement said. Mr. Lai, the president, said the Taiwanese government was concerned about the global impact of the tariffs.

The Taiwanese government “thought too optimistically about the relationship with Trump,” said Jason Hsu, a senior fellow at Hudson Institute and former member of Taiwan’s legislature for the opposition Nationalist Party. “It was a little bit too naïve thinking Trump would be nice to them, especially after the TSMC announcement.”

Chris Buckley contributed reporting from Taipei, Taiwan.