Wednesday Rundown
Keeping you up to date on pressing news from May 5 – 7.
Welcome back to Ghost Reader – a publication to help you stay informed when you don’t have the time, access or energy to keep pace with today’s news cycle.
This edition is jam-packed - here’s what you can look forward to:
High-Impact Events: The latest conflict between India and Pakistan, as well as how the Trump administration is defending access to an abortion pill. We’ll also highlight the REAL ID requirement that takes effect today.
Executive Orders & Legislative Impacts: A Supreme Court ruling on transgender people serving in the military, proposed increase to the small business loan limit, and contradictory U.S. intel to President Trump’s invocation of the Alien Enemies Act to deport Venezuelan migrants.
Human Experience & Rights: More news related to the intersection of humans and artificial intelligence (AI), including AI bot “hallucinations” and what to know about privacy when using Meta’s new AI chatbot.
Federal Government Operations: The latest on how the ‘Department of Government Efficiency’ (DOGE) is pooling sensitive data across federal agencies.
Good for the Soul: A look at negativity bias and how to combat it.
Let’s get into it.
High-Impact Events
Catch me up on what’s happening between India & Pakistan
Early Wednesday, India attacked several parts of Pakistan and Pakistani-controlled territory in an escalation of longstanding tensions between the countries.
The strikes were largely in response to an April 22 terrorist attack on the Indian-controlled side of Kashmir that left at least 26 tourists dead and more than a dozen injured. India suggested Pakistan was behind the attack, while Pakistan repeatedly denied the claim.
What should I know?
The historical context: This is the latest in a decades-long conflict over Kashmir – a scenic valley in the Himalayas that sits between India and Pakistan. For the sake of conserving newsletter space, check out a brief timeline of back and forth conflicts over the territory since 1947 via this gift link from The New York Times.
Where things stand today: The April 22 attack was one of the worst on Indian civilians in decades. Almost immediately, Indian officials suggested Pakistan was involved and vowed severe punishment in retaliation. Wednesday, India struck nine Pakistani sites and Pakistan’s side of Kashmir after it said evidence “pointed toward the clear involvement of Pakistan-based terrorists” in the April attack. As of late Wednesday, the death toll in Pakistan reached 31, while at least seven people died in Indian-controlled Kashmir.
What might happen next: Pakistan responded to Wednesday’s attacks by calling them “an unprovoked and blatant act of war.” It also said the strikes “would not go unanswered.” With Pakistan and India’s history of war, what happens next between these two nuclear-armed countries could be significant.
How the White House responded: When asked by a reporter about the attack, President Trump said it was “a shame” and he “hopes it ends very quickly.” Secretary of State Marco Rubio said he is monitoring the situation and Washington would continue to engage the countries to reach a “peaceful resolution.” Both India and Pakistan are partners to the U.S.
Tell me about the abortion pill, mifepristone
Monday, the Trump administration pushed forward in defending access to the abortion pill mifepristone, in line with a legal challenge that began during former President Joe Biden’s administration.
What should I know?
What mifepristone is: A drug used in combination with another medication called misoprostol to end a pregnancy that is less than 70 days old. It’s used in more than 60% of U.S. abortions.
The context: In recent years, there were multiple lawsuits related to restrictions around the abortion pill. Most recently, Missouri, Kansas, and Idaho filed a lawsuit claiming the U.S. Food and Drug Administration (FDA) acted improperly in 2016 and 2021 when it eased restrictions on the pill, including allowing it to be prescribed via tele-health and dispensed by mail. Last year, the U.S. Supreme Court rejected a bid by anti-abortion groups and doctors to restrict access to the drug, finding they lacked legal standing to challenge the FDA regulations.
Where things stand now: This week, the Department of Justice filed a brief in the Texas federal court urging the judge to dismiss the lawsuit on procedural grounds. The DOJ lawyers said Texas is not the proper venue for the lawsuit and the states lack standing because they are not harmed by the regulations. In response, the states argued their Medicaid health insurance programs would have to pay to treat patients who suffer complications from using mifepristone. They also argued they should be allowed to remain in Texas (where the original lawsuit was filed) because it would be inefficient to send the case to another court after more than two years of litigation.
The Trump administration’s approach: While campaigning last year, Trump said he did not plan to ban or restrict access to mifepristone. In February, Secretary of Health and Human Services Robert F. Kennedy Jr. told Fox News the president asked for a study on the safety of abortion pills and had not decided on whether to tighten restrictions on them.
20 years later, the REAL ID deadline is here
After years of delays, the REAL ID deadline that requires individuals to present a REAL ID or other form of acceptable documentation to board flights finally took effect.
What should I know?
The background: Congress passed the Real ID Act in 2005 to increase aviation security following the September 11, 2001 terrorist attacks. The deadline for implementation was repeatedly delayed since its original 2008 target date.
What a REAL ID is: In general, it’s a security-enhanced driver’s license or state identification card. In most states, it looks like a standard driver’s license with a star in the upper right corner.
Who needs a REAL ID: People who plan to board commercial flights in the U.S. or access certain federal facilities if they don’t have other forms of acceptable identification.
How to obtain a REAL ID: Through the Department of Motor Vehicles (DMV). While requirements differ slightly by state, all applicants must provide documentation showing their full legal name and date of birth, proof of address, and proof of lawful status in the U.S. Applicants must also bring a Social Security card, SSA-1099 form, or document that shows their name and Social Security number, such as a pay stub or W-2.
Some stats: According to estimates from the think tank, Center for American Progress, as of 2024, nearly 47% of Americans lack a valid passport. The states with the largest number of citizens lacking a passport are West Virginia (79.3%), Mississippi (77.9%), and Alabama (72.3%). The states with the smallest numbers of citizens that don’t have valid passports are New Jersey (20.1%), California (28.2%), and Massachusetts (28.8%).
What happens if you don’t have a REAL ID and need to fly: There are about a dozen alternative forms of identification that can be used.
Other forms of acceptable identification: They include a valid passport or passport card, Department of Homeland Security-issued trusted traveler card (such as Global Entry, NEXUS, SENTRI, FAST), U.S. Department of Defense ID (such as a military ID), U.S. permanent resident card, border crossing card, Tribal Nation photo ID, HSPD-12 PIV Card (issued to federal employees and contractors), foreign government-issued passport, Canadian provincial driver’s license or Indian and Northern Affairs Canada card, transportation worker ID credential, U.S. Citizenship and Immigration Services Employment Authorization Card, U.S. merchant mariner credential, or Veteran Health ID card.
What DHS and TSA are saying: Tuesday, Secretary of Homeland Security Kristi Noem told a congressional panel that people will still be allowed to fly if they don’t have the REAL ID, but they may be diverted to a different TSA line with extra steps (and time) to get through security. A Transportation Security Administration (TSA) official told Forbes that TSA officers will look for alternative ways to help people prove their identity, saying “We are used to dealing with people who lose their wallets and IDs, and we have ways to vet passengers and make sure we’re satisfied the person standing in front of us is the person they say they are.”
Executive Orders & Legislative Impacts
Let’s talk about the ruling on trans people in the military
Tuesday, the Supreme Court paused an order by a federal judge in Washington state that barred the government from implementing a policy that would prohibit transgender people from serving in the U.S. military.
What should I know?
The context: In 2021, former President Joe Biden signed an executive order that allowed transgender people to serve openly in the U.S. military. On President Trump’s first day in office this year, he revoked that order and issued a different one requiring Secretary of Defense Pete Hegseth to ban “individuals with gender dysphoria.” On February 26, the Department of Defense issued that ban, which generally disqualifies anyone who has gender dysphoria or undergone medical interventions to treat gender dysphoria from serving in the military.
The rationale: The Defense Department said the “medical, surgical, and mental health constraints on individuals who have a current diagnosis or history of, or exhibit symptoms with, gender dysphoria are incompatible with the high mental and physical standards necessary for military service.”
What “gender dysphoria” is: The medical term for the psychological distress someone experiences when their gender identity differs from the sex they were assigned at birth.
About the lawsuit: Following the Defense Department ban, seven current transgender members of the military, one transgender person who would like to join the military, and a nonprofit with members who are either transgender troops or would like to be, challenged the new policy in federal court. The presiding judge agreed with the plaintiffs that the ban violated the Constitution’s equal protection guarantee and was essentially a “de facto blanket ban on transgender service.”
How it got to the Supreme Court: An appeals court rejected the government’s request to freeze the order, leading the Trump administration to ask the Supreme Court to rule on April 24. In this, the Trump administration told the Supreme Court that without its intervention, the lower court’s order would stay in place while litigation continues in the appeals process. The Trump administration said it would be “too long for the military to be forced to maintain a policy that it has determined, in its professional judgment, to be contrary to military readiness and the Nation’s interests.” In response, the plaintiffs said if the Supreme Court paused the order, it would allow the government to begin discharging thousands of transgender service members, “ending distinguished careers and gouging holes in military units.”
What the Supreme Court decided: In a 6-3 decision, it granted the Trump administration’s request to pause the order while the appeals process plays out. As is the case for emergency appeals, the court did not provide an explanation for its decision.
What it means: The Defense Department policy can take effect for now, with the impacts on current transgender military members to be seen.
I heard the small business loan limit might increase
On May 2, three members of Congress introduced the “Made in America Manufacturing Finance Act,” which would raise the current $5 million cap on Small Business Administration (SBA) 7(a) and 504 loans to $10 million for manufacturers.
What should I know?
Who introduced the bill: House Rep. Roger Williams (R-TX), Senator Joni Ernst (R-IA), and Senator Chris Coons (D-DE).
About the programs included: The 7(a) program offers flexible financing for working capital, equipment, and real estate. The 504 program is typically used for fixed assets like land and large machinery. Both are made by private lenders but backed by the SBA.
Some context: The current $5 million limit hasn’t changed since 2010. If it had been adjusted for inflation, the limit would be about $7.5 million today.
The SBA endorsed the bill: SBA Administrator Kelly Loeffler called the bill a way to help manufacturers grow and compete. She said the bill fits with a broader Trump administration push to bring manufacturing back to the U.S., paired with tariffs meant to reduce dependence on foreign production.
What the broader landscape says: While there is support, many question its focus on manufacturing while some hope it might pave the way for other sectors to be included. Chris Hurn, founder of Phoenix Lender Services, said the bill is a good first step and chance to spotlight SBA spending, but limiting it to manufacturers is too narrow. Similarly, Matthias Smith, founder of Pioneer Capital Advisory, said he hasn’t seen much demand for manufacturing acquisitions, in part because they are hard to finance and require deep industry knowledge. He views the bill as more symbolic than practical, especially as the SBA is set to roll back pandemic-era rules in favor of stricter credit checks, higher equity requirements, and tougher underwriting standards.
Catch me up on intel that went against Trump’s deportation rationale
A recently released memo shows U.S. intelligence agencies concluded last month the Venezuelan government does not direct the Tren de Aragua gang’s activities, essentially contradicting the rationale President Trump used in invoking the Alien Enemies Act to deport Venezuelan migrants to a prison in El Salvador as a matter of “national security.”
What should I know?
A reminder: When Trump invoked the Alien Enemies Act, he declared the U.S. was under invasion by a hostile foreign nation and, therefore, suspected gang members – in this case, Tren de Aragua – could be rapidly deported without judicial review. This led to three flights of alleged Venezuelan gang members being deported to El Salvador, prompting what has become a flurry of legal action against the deportations.
About the intel: The National Intelligence Council’s “sense of the community” memo found that while Tren de Aragua likely takes advantage of widespread lawlessness under Venezuelan President Nicolas Maduro, there are no direct ties to the government.
What else the intelligence agencies found: They challenged Trump’s depiction of Tren de Aragua as a “preeminent threat” to national security, or even much of a criminal organization, at all. The memo said Tren de Aragua is small, with a focus on low skill criminal activity and a decentralized structure that makes it highly unlikely they’re coordinated enough to do large volumes of human trafficking and migrant smuggling.
What this means: It’s more evidence that will likely be pulled into the many lawsuits facing the Trump administration over its use of the Alien Enemies Act and deportation of hundreds of individuals to El Salvador without due process; including some who are legal U.S. citizens, have no criminal history, and no connection to any gang activity.
Human Experience & Rights
Tell me about AI’s error problem
More than two years after the creation of ChatGPT, humans are increasingly using artificial intelligence (AI) bots for daily tasks. The newest and most powerful technologies – so-called reasoning systems – are generating an increasing number of errors, though. While their math skills have improved, their handle on facts is worse.
What should I know?
Some context: Today’s AI bots are based on complex mathematical systems that learn skills by analyzing enormous amounts of data. They do not – and cannot – decide what is true or false. Due to this, they sometimes make things up – a phenomenon AI researchers call “hallucinations.”
About these “hallucinations”: In one recent test, the hallucination rates of newer AI systems were as high as 79%. Examples might include asking an AI bot for a good marathon on the West coast and it returning a suggestion for a race in Philadelphia (on the East coast). Or when asking a question, it might cite a source that doesn’t include the information it’s responding with. While hallucinations like this might not be a big problem, they can become a serious issue for anyone using the technology with legal documents, medical information, or sensitive business data.
What’s happening now: Companies like OpenAI and Google have worked to improve their AI systems and reduce the frequency of these errors for years. But with the introduction of new reasoning systems, the errors are back – and rising. According to a recent report by OpenAI, the company’s latest systems hallucinate at a higher rate than the previous system.
Why it’s so hard to pinpoint the “why”: Hannaneh Hajishirzi, a University of Washington professor and researcher with the Allen Institute for Artificial Intelligence, is part of a team that recently devised a way to trace a system’s behavior back to individual pieces of data it was trained on. Because the systems learn so much data, though – and because they can generate almost anything – even a new tool like this can’t explain exactly how the models work. Another issue is that reasoning models are designed to “think” through complex problems before settling on an answer. As they do, they run the risk of hallucinating throughout the process, with the errors compounding the more time they spend “thinking.”
What this means for an average user: To leverage tips for using AI safely, continue exercising a healthy level of skepticism, and always do your own fact-checking. Though these tools are helpful in some situations – like writing papers, summarizing office documents, generating code, or making slide decks – their mistakes can lead to problems; and often, the AI doesn’t know it made a mistake.
Speaking of … let’s talk about Meta’s new AI app some are calling “creepy”
Last week, Meta launched the Meta AI app – a home for the company’s AI chatbot that promises users a more “personalized” AI experience with tailored answers and advice. It also includes a social network for people to share their AI conversations and images.
What should I know?
Some important things if you plan to use the app: There are five main privacy warnings you should know – check them out below.
“Sharing” is public: There is a “share” button on the top right corner of every Meta AI chat. Tapping will post the chat contents to a public “Discover” tab. This means anything someone shares is public – there is no option to share with only one’s Facebook or Instagram friends.
It’s linked to Facebook and Instagram: If you create a profile for the app using your existing Facebook or Instagram accounts, the AI will have access to information from your profile. Meta AI can also share information derived from your chats back with Facebook and Instagram. If you don’t want that data mingling, you’ll have to create a separate Meta AI account.
It remembers everything: Meta AI keeps a transcript or voice recording of every conversation you have with it. The AI learns from what you ask and can reference it in future chats. It also keeps a list of what it deems to be key facts about you in a separate “Memory” file, that you can view in the app’s Settings. While you can delete past chats and “Memories,” you can’t stop Meta AI from recording them.
Your chats are training Meta’s AI: The contents of your chats – words, photos, emojis, voice – are fed back into Meta’s AI training systems. Meta even warns in its service terms to not share information you don’t want the AI to use and retain.
Your chats can be used for ads: While Meta AI doesn’t currently include ads, CEO Mark Zuckerberg has hinted they’re coming.
Some examples of how all this plays out: This week, a reporter from The Washington Post tested the app out. Here are anecdotes from what he experienced:
When he first opened the app, he asked it to describe him in three emojis. Meta AI could, by drawing on years of personal information tracked via Facebook and Instagram.
He chatted with Meta AI about some intentionally sensitive topics. Afterward, it built a “Memory” file about him capturing the fake topics he brought up.
Meta AI kept a copy of everything he said. The reporter was eventually able to delete the “Memory” file and conversations, but it required a lot of research and digging into deeper levels of the app’s Settings.
Why this matters: Meta AI is is pushing the limits on privacy in ways that go further than rivals ChatGPT from OpenAI or Gemini from Google. While the reporter’s conversations were fictional tests, an increasing number of people are turning to AI for personal matters. Meta AI’s personalization brings a new set of privacy risks people might not yet appreciate, let alone understand or know how to manage.
What an expert said: Ben Winters, director of AI and data privacy at the Consumer Federation of America, said Meta AI’s disclosure and consumer choices around privacy settings are “laughably bad” and he would only use the app for surface-level, fun prompts or things that don’t have anything to do with one’s personal information.
Federal Government Operations
It’s been a while … let’s talk about DOGE and data
As mentioned in previous Ghost Reader editions, the so-called ‘Department of Government Efficiency’ (DOGE) is racing to build a centralized database that pulls together data from across many federal agencies. Many government workers say this “mission” often violates or disregards core privacy and security protections meant to keep U.S. citizens’ personal information safe and secure.
What should I know?
The “why”: The effort aims to support Trump administration priorities, such as finding and deporting undocumented immigrants and rooting out “waste, fraud, and abuse” in the federal government. It also follows a March executive order to eliminate “information silos” across federal agencies.
What’s happening: Federal workers say DOGE officials have sought to merge databases that were long kept separate. In some cases, DOGE has removed protections around sensitive information such as Social Security numbers, birth dates, employment history, disability records, medical documentation, and more.
How this is different from what normally happens: Typically, data sharing in the federal government requires multiple steps. It includes legislative permission, public notices of what the government is doing, and “computer matching” agreements between agencies specifying what is to be shared and why. Independent inspectors general also ensure information is used properly. According to federal workers and government officials across 10 federal agencies speaking under anonymity, none of these steps have been followed consistently by the DOGE team.
Potential implications: The centralization effort and fast pace increase the risk of exposing data to hackers, according to security analysts. Civil rights advocates also worry the data could be used against political foes or for targeted decisions about funding or basic government services. In response to what’s happening, Albert Fox Cahn, executive director of the Surveillance Technology Oversight Project, said, “This threat isn’t just going to exist tomorrow – it’s going to exist for decades to come.”
Good for the Soul
As humans, it’s easier to focus on the negative than the positive. Let’s explore why and strategies for overcoming what’s known as “negativity bias.”
What should I know?
About “negativity bias”: It’s our cognitive tendency to attend to, learn from, and use negative information more than positive information. It can explain why we often recall and think about insults more than compliments or dwell on unpleasant or traumatic events. For some, negativity bias can be correlated with increased stress, depression, and anxiety.
Where it comes from: Negativity bias is important to human functioning since it helps protect us from harm. While that may look differently today than it did centuries ago, the tendency is thought to be an adaptive evolutionary function. Thousands of years ago, our ancestors were exposed to immediate environmental threats. For them, being more attentive to negative stimuli often played an important role in their survival.
There are some upsides: While negativity is often viewed – well, negatively – it can sometimes be useful, as negative feelings contain important information that can inform growth and future decisions. For instance, focusing on negative feelings can spur us to solve problems or do something differently. It can also push us to action.
Ways to turn negativity bias into productive action:
Try switching your attention to the positive: When our brains default to the negative, we first need to notice it. Then, we can try to correct it or actively shift our attention to something good.
Make gratitude a habit: Set a time each day to write down or reflect on a handful of things you’re thankful for. Naming what you appreciate, or the “good” in your day, helps train your brain in a new habit. While it’s hard at first, the more you do it, the easier and more automatic it gets, which can boost overall well-being.
Use the negative to break bad habits: If you want to change a bad habit, lean into negativity bias and think about negatives associated with the habit. According to research, increasing how badly we feel in a situation is more effective in creating behavioral change. So by focusing on the negative consequences or outcomes of a habit you want to change, it can help decrease the behavior over time.
Connect with people in a constructive way: Focusing on the negative can lead us to be self-focused. Finding ways to connect with and help others can be a shortcut to breaking out of the negativity spiral. Channeling negative feelings into positive action can not only make a difference in the community, but also leave us feeling better as a result.
So while we as humans have a tendency to focus more on the bad than good – and that might be especially true in today’s world – there are strategies to help disrupt the bias that can lead to greater overall well-being. So the next time you find yourself zeroing in on the negative and overlooking the positive, give one of them a try.
That’s all for today. Until Friday, let’s keep working to create clarity in the chaos.
Did you find this newsletter helpful? If so, please subscribe and share.