News
WASHINGTON, DC, USA –– White House Press Secretary Karoline Leavitt answers questions from journalists on Jan. 28, 2025.
WASHINGTON — As the Trump administration increasingly welcomes vloggers and social media influencers into press briefings and the Oval Office, established outlets like the Associated Press find themselves excluded from the century-old press pool, sparking controversy about what "transparency" truly means.
Joshua Sukoff is a photojournalist from Long Island, New York. He is currently studying journalism at Northwestern University.
Young adult woman accessing artificial intelligence, managing digital information, and exploring future technology within an office setting
The harder question is how many capable people organizations will need. My latest lesson about artificial intelligence did not come from a research report or technology conference. It came from running the Security and Sustainability Forum.
Until recently, I depended heavily on an assistant to manage much of the website and webinar production process. Now I can do many of those tasks myself with Claude and ChatGPT. They help me develop content, revise web pages, solve technical problems, and manage projects that once required considerably more human support.
I am not eliminating my assistant’s position. I am redefining it. Instead of spending as much time on routine production, my assistant can concentrate on creative design and technical review. That work requires human judgment and can help SSF improve its programs and grow.
That experience led me back to an idea Apple popularized: the “directly responsible individual,” or DRI. Under Steve Jobs, an effective meeting reportedly ended with an action list, and each action carried the name of the person responsible for ensuring it was completed.
The AI-enabled DRI could become a defining worker of the future. One person will be able to direct artificial capabilities for research, writing, coding, design, and analysis. The individual will define the objective, divide the problem into parts, judge the results, and take responsibility for the final product.
That person need not be an engineer or computer scientist. A humanities major who understands context, formulates good questions, detects weak reasoning, and communicates clearly could be well positioned. AI can increasingly perform technical production. Human value may shift toward determining what should be done, recognizing when the result is wrong, and accepting responsibility for the outcome.
These could be fascinating jobs. My concern is not whether people can prepare for them. It is the number of such people an organization will need.
We tell workers to learn AI, and they should. But if everyone becomes more productive, it does not follow that companies will need everyone. AI could give more people access to sophisticated work while reducing the number of positions available to them.
Organizations will face a choice. They can use AI to produce yesterday’s output with fewer employees, or they can redirect the capacity it releases toward better products, new services, stronger relationships, and growth. Before eliminating a position, a leader should ask what higher-value work that person could now perform.
There is another problem. The future DRI will need judgment, but judgment develops through experience. Junior professionals learn by researching, drafting, analyzing, making mistakes, and receiving criticism. AI is absorbing much of that work. We may be creating a workplace that demands more judgment while eliminating the work through which judgment has traditionally been acquired.
AI also differs from earlier technologies. It acts upon the intellectual work through which we create new products, discoveries, and technologies. AI companies are already using it to help develop more capable AI, although full autonomous self-improvement has not arrived and may never arrive. The tool is beginning to participate in the process that improves the tool.
There is no realistic way to stop this change. The competitive advantages are too great. People will need curiosity, judgment, imagination, problem-definition skills, and the willingness to be responsible for results. Organizations will have to decide whether AI’s gains mean fewer jobs or greater ambition.
But preparation cannot rest entirely on individuals and employers. If the economy eventually requires much less human labor, telling everyone to master AI will not answer the larger question. Society will have to decide how opportunity, income, and purpose are distributed when productive capacity grows faster than the demand for people.
We do not yet know whether AI will take us that far. We know that it has started moving us in that direction, and that each generation of the technology may accelerate the next.
The Industrial Revolution showed how profoundly technology could reorganize society. It may have been the appetizer. AI is the main course.
Edward Saltzberg is the Executive Director of the Security and Sustainability Forum. He is a Visiting Scholar at George Washington University and writes the Stability Brief.
AI résumé screening can filter out qualified candidates before a human ever looks. Here's why transparency and human oversight in hiring matter.
A job seeker can spend hours tailoring a résumé, checking qualifications and writing a thoughtful application, believing the next step will be a person deciding whether to schedule an interview.
Increasingly, that assumption may be wrong.
Before a recruiter reads a résumé, software may sort applications, identify keywords, rank candidates, answer screening questions or determine which applicants advance. Some systems are relatively simple filtering tools. Others use machine learning or artificial intelligence to evaluate applicants across multiple data points.
There is nothing inherently wrong with using technology to manage a hiring process. Employers have legitimate reasons to use it. A company receiving hundreds or thousands of applications cannot expect recruiters to manually examine every document with unlimited time and resources.
The harder question is what happens when efficiency becomes a substitute for judgment.
When an automated system influences who receives a chance to compete for a job, it is no longer merely an administrative convenience. It becomes part of the infrastructure through which people gain access to economic opportunity.
That deserves more scrutiny.
The first decision may happen before a human sees the candidate
The U.S. Equal Employment Opportunity Commission has examined the benefits and risks of artificial intelligence and automated systems in employment decisions, including recruitment and hiring. The agency has emphasized that existing employment-discrimination laws continue to apply when employers use automated technologies.
That distinction matters.
An algorithm does not need to explicitly discriminate against someone to produce a problematic outcome. A screening system can rely on criteria that appear neutral but function differently for different applicants.
Consider a candidate who took several years away from paid employment to care for a parent. Consider another who built skills through military service, community work or self-employment rather than through a conventional career path. Or consider someone whose experience is highly relevant but described using terminology that differs from the language used in a job posting.
A human recruiter might recognize the connection.
A rigid screening rule might not.
In testimony before the EEOC, ReNika Moore discussed how automated screening tools can filter or rank applicants using knockout questions, keyword requirements, specific qualifications and other criteria. Her testimony also raised concerns about systems that may screen for employment gaps or narrow descriptions of experience in ways that can exclude qualified candidates.
These are not arguments against technology. They are arguments against assuming that whatever is measurable is necessarily what matters.
A résumé is not a person
This is perhaps the most important limitation of automated résumé screening.
A résumé is an imperfect representation of a human being. It contains selected facts about someone's education, work history, skills and accomplishments. It does not contain the whole person.
Years of experience do not automatically equal competence. A particular job title does not guarantee ability. A missing keyword does not prove that someone lacks a skill.
Even sophisticated systems face a fundamental problem: predicting someone's suitability for a job from indirect signals.
The EEOC's examination of automated employment systems has highlighted concerns about screening tools that can reject candidates before human review, as well as systems that rely on highly specific credentials, job-experience descriptions or other rigid criteria.
That should make employers cautious about treating an algorithmic ranking as an objective measurement of merit.
A system can be consistent and still be wrong.
It can process thousands of résumés in seconds and still miss the person who would have been the best hire.
But the human process is hardly perfect
There is an understandable temptation to frame this as a choice between biased technology and unbiased human beings.
That would be a mistake.
Human hiring decisions are not automatically fair. Recruiters and hiring managers have their own assumptions, preferences and blind spots. Humans can overlook qualified candidates because of familiarity with certain employers, conventional career paths, educational credentials or simple fatigue after reviewing hundreds of applications.
Automation can sometimes reduce certain inconsistencies. It can help recruiters handle enormous application volumes and draw attention to qualifications that might otherwise be overlooked.
That is why the debate should not be “AI versus humans.”
The more useful question is: Where should technology make decisions, and where should humans remain accountable for them?
That is a much harder question, but it is also the one employers need to answer.
Efficiency should not become an excuse for opacity
The most troubling development would be a hiring system in which nobody can clearly explain why an applicant was screened out.
Job seekers generally cannot see how a company's hiring technology weighs their experience. They may not know whether a résumé was evaluated by a keyword filter, an automated scoring system, an assessment tool or a combination of technologies.
That creates an asymmetry of power.
The employer knows how the gate works. The applicant is simply told whether the gate opened.
Some jurisdictions have already begun addressing this issue. New York City's Local Law 144 requires covered employers and employment agencies using automated employment decision tools to satisfy requirements involving bias audits, public information about those audits and notice to workers or candidates.
Whatever one thinks about the details of such rules, the underlying principle is worth considering nationally: people should not be left completely in the dark when automated systems play a significant role in determining access to employment.
Transparency does not necessarily mean publishing an algorithm's source code.
It can mean something much simpler.
Was an automated tool used? What kind of decision did it influence? Was there a meaningful opportunity for human review? What happens when the system produces an obviously questionable result? How can an applicant request an accommodation or raise a concern?
Those questions are not anti-technology.
They are basic questions of accountability.
The burden cannot fall entirely on job seekers
There is already a growing expectation that applicants must learn how to “beat the algorithm.”
That phrase reveals something important about the system.
When people are encouraged to load their résumés with keywords, mirror the wording of job descriptions or restructure their professional histories primarily to satisfy an automated filter, the hiring process risks becoming a competition in algorithmic optimization rather than an assessment of who can actually perform the job.
Job seekers should certainly present their qualifications clearly. But they should not have to become amateur software engineers merely to ensure that a qualified human gets a chance to evaluate them.
The responsibility belongs elsewhere, too.
Employers should understand what their automated tools actually measure, validate whether those measurements relate to legitimate job requirements and regularly examine outcomes for evidence of unintended exclusion.
Technology vendors should be candid about limitations and give employers meaningful ways to monitor performance.
Recruiters and hiring managers should retain enough authority to question automated recommendations rather than treating them as final judgments.
And policymakers should focus on practical protections around accountability, notice, accessibility and discrimination without assuming that every form of workplace automation requires exactly the same regulation.
Human oversight should mean more than a human somewhere in the process
There is a difference between human involvement and human judgment.
A recruiter who automatically accepts an algorithm's ranking is technically part of a human-supervised process. But that does not necessarily provide meaningful oversight.
Real oversight means people can investigate unusual outcomes, challenge automated recommendations and reconsider candidates who may have been filtered out for reasons unrelated to their ability to perform the job.
This becomes particularly important when automated systems encounter candidates whose backgrounds do not fit conventional patterns.
Career paths are becoming less linear. People change industries. They return to work after caregiving. They acquire skills outside traditional degree programs. Veterans translate military experience into civilian occupations. Workers rebuild careers after layoffs. Immigrants bring professional experience from other countries.
A hiring system designed around yesterday's career path may have difficulty recognizing tomorrow's worker.
That is a governance problem as much as a technology problem.
The goal should be better hiring, not simply faster hiring
The pressure to make hiring faster is understandable.
But speed is only one measure of a successful hiring system.
A process can be fast, inexpensive and highly automated while still producing poor matches or excluding people who should have received consideration.
The better standard is whether technology helps employers identify qualified people while preserving fairness, accountability and meaningful human judgment.
AI may make parts of recruitment better. It can reduce administrative burdens, help employers process large applicant pools and potentially help recruiters focus more of their time on candidates rather than paperwork.
But realizing that promise requires a basic principle:
An automated hiring system should assist the decision-maker without becoming an invisible decision-maker.
The stakes extend beyond individual résumés.
Employment is one of the primary ways Americans gain economic security, independence and a sense of participation in society. When access to employment is increasingly mediated by technology, the design and governance of that technology becomes a public-interest issue.
The question is therefore not whether artificial intelligence deserves a place in hiring.
It does.
The question is whether the people affected by it will have a fair shot at being seen.
A résumé should open the door to consideration, not become the reason the door never opens.
Michelle Brenier is a SaaS and technology content writer specializing in AI, recruitment technology, emerging technologies and the changing nature of work. He contributes content expertise to Jump Resume Builder, where he focuses on career technology and tools that help job seekers navigate the modern hiring process.
U.S. President Donald Trump answers questions from reporters before boarding Air Force One on August 21, 2026 in Joint Base Andrews, Maryland.
The principle is simple. Just like politicians are not supposed to pick their voters, government officials don’t get to pick the reporters who cover them.
If they did, the First Amendment would be meaningless. and we would no longer live in anything resembling a constitutional republic. That is why President Trump’s September 18 announcement that he is banning reporters from CNN, MSNow, and Politico from the White House is so dangerous.
Whatever one thinks of the quality of their news coverage, all of us have a stake in resisting the president’s brazen attempt to silence and punish news outlets who refuse to bend a knee to him. Because it is a form of viewpoint discrimination, the ban violates the First Amendment, and I expect that the courts will soon say so.
But whatever the courts do, other members of the Whiter House Press Corps should pack their bags and reuse to give President Trump the coverage he so clearly craves.
The American public learned of the president’s intention from a post on Truth Social. “I am proud to announce that, effective immediately, I am banning Fake News CNN, MSNOW (who recently changed their name from MSNBC due to lack of viewership and credibility!), and Politico (The recipients of an illegal and ridiculous $8 Million Dollar subscription, an All Time Record, directly from the United States Government, under Crooked Joe Biden, in order to keep them “alive.” Seems like corruption to me!), from the White House as a result of their constant “reporting” FAKE NEWS!”
“Media Outlets,’ he continued, “shouldn’t be able to constantly write or report FICTION and LIES when they’re covering the President of the United States, the Trump Administration, or the United States of America. Other Fake News Media Outlets to follow.”
Actually, the First Amendment gives broad scope to the press to publish false information so long as they don’t do it with “actual malice.” In its 1964 decision in New York Times v. Sullivan a unanimous Supreme Court held that “When a statement concerns a public figure… it is not enough to show that it is false for the press to be liable for libel. Instead, the target of the statement must show that it was made with knowledge of or reckless disregard for its falsity.”
The president hates that decision and would love to see it overturned
As Justice William Brennan noted, “In the effort to persuade others to his own point of view, the pleader, as we know, at times resorts to exaggeration, to vilification of men who have been, or are, prominent in church or state, and even to false statement. But the people of this nation have ordained, in the light of history, that, in spite of the probability of excesses and abuses, these liberties are, in the long view, essential to enlightened opinion and right conduct on the part of the citizens of a democracy."
“That erroneous statement,” Brennan observed, “is inevitable in free debate, and that it must be protected if the freedoms of expression are to have the ‘breathing space’ that they ‘need . . . to survive.’"
The president hates that decision and would love to see it overturned.
And, in an Oval Office meeting with reporters after his Truth Social Post, the president made clear that what irked him about reporters from CNN, MsNow and Politico was that they are “people to write purposely negative stories….”
While the welcome mat is out for reporters who write positive stories about him and his administration. If a reporter does not toe the line, then “that’s fine, but I don’t have to let them into my — into the people’s house.”
Put aside the telling slip of the tongue and what remains is a boldfaced assertion of a presidential prerogative to punish news outlets because of what they say and the viewpoint they express. There is nothing more central to the First Amendment than that.
Indicators of viewpoint discrimination include official action that “singles out particular speakers may be aimed at restricting certain content or certain viewpoints… (or that) restricts only a particular form of expression may be aimed at suppressing a particular viewpoint.” Excluding three news outlets because of their “negative coverage” is about as clear an indication of viewpoint discrimination as one could get.
As the famous Judge Learned Hand once explained “The First Amendment presupposes that right conclusions are more likely to be gathered out of a multitude of tongues than through any kind of authoritative selection. To many, this is, and always will be, folly, but we have staked upon it our all."
The president and others in his administration have made it abundant clear that they have no interest in hearing a “multitude of tongues.”
And this is not the first time the president has tried to exclude a reporter whose coverage he did not like. In November 2018, closed to the halfway point of his first term his White House suspended the press credentials of CNN’s Jim Acosta after Acosta “challenged the president on his characterization of the migrant caravan moving through Mexico as an ‘invasion.’"
CNN sued and Federal District Judge Timothy Kelly ordered the White House to restore Acosta’s credentials on the ground that it hard deprived him of due process by taking it away in a summary fashion.
He based his ruling on a 1977 decision by the Court of Appeals for the D.C. Circuit. As that court explained, “denial of a White House press pass to a bona fide journalist violates the first amendment unless it furthers a compelling governmental interest identified by narrowly and specifically drawn standards…(and) that denial of a White House press pass constitutes a deprivation of "liberty" without due process of law within the meaning of the fifth amendment because it interferes with the free exercise of the profession of journalism.”
It added, “notice, opportunity to rebut, and a written decision are required because the denial of a pass potentially infringes upon first amendment guarantees. Such impairment of this interest cannot be permitted to occur in the absence of adequate procedural due process.”
Apparently, the president has not learned his lesson. Or maybe he thinks that when he took the oath of office, he became the decider who could do whatever he wants to whoever he wants.
It looks like he will again have to be disabused of that notion in a court of law and hopefully by the solidarity of the White House Press Corps. But ultimately it will be up to voters to send the message that he is the president in a constitutional republic, not a monarch.
In the meantime, all of us should remember, even if the president does not, Thomas Jefferson’s admonition that "(O)ur liberty depends on the freedom of the press, and that cannot be limited without being lost… (and that) were it left to me to decide whether we should have a government without newspapers or newspapers without a government, I should not hesitate a moment to prefer the latter.”
Austin Sarat is the William Nelson Cromwell professor of jurisprudence and political science at Amherst College.
A 2026 study found AI beat elite debaters, until it was slowed to human speed. Why persuasion throughput matters and how to make chatbot tempo transparent.
A political chatbot can answer before a citizen has finished reading its previous message. It can add statistics, historical examples, causal claims, and links at a rate no human campaign volunteer could match. That does not necessarily make its case wiser. It changes the contest by allowing one speaker to control the tempo.
A June 2026 preprint offers unusually direct evidence of this problem. Across four preregistered experiments involving 18,978 conversations from 6,923 people, frontier AI systems were more persuasive than laypeople, tournament winners, professional canvassers and elite competitive debaters. The human experts were not casual opponents: the study included world and continental champions, paid preparation, live practice and substantial performance incentives.
The most revealing result came when the researchers changed the rules of the exchange. Elite debaters had produced replies averaging 54 words after roughly 95 seconds. The AI produced about 294 words with sub-second latency. When the system was constrained to roughly human-length messages and human writing speed, its advantage over coached elite debaters fell to a statistically indistinguishable zero.
The study is a preprint, not a final verdict, and its experiments were conducted in controlled text conversations in the United Kingdom. It does not prove that any chatbot will decide an election. It does, however, identify a democratic variable that is largely missing from public debate: persuasion throughput, or the amount of argumentative material a system can deliver within a citizen's limited attention.
Our current safeguards focus on important questions. Was the message generated by a machine? Who paid for it? Was personal data used to target the recipient? Is the image or voice synthetic? Those questions address identity and targeting, but they do not address what happens after a person enters the conversation.
A chatbot can present dozens of claims before the user has inspected the first source. In the 2026 study, unconstrained AI produced about 37 fact-checkable claims per conversation; the human-paced version produced about 12. Across conditions, fact density strongly predicted persuasive impact. The advantage was not simply that the machine sounded more empathetic or more human. It could place more information on the table before its counterpart could answer.
A peer-reviewed Science article published in December 2025 points in the same direction from another angle. In experiments with 76,977 participants, 19 models and 707 political issues, post-training and prompting designed for persuasion mattered more than personalization or model scale. The same interventions that increased persuasion also reduced factual accuracy. Speed and density can therefore amplify useful information and error at the same time.
The appropriate response is not a government word limit on political speech. It is to make the tempo of machine persuasion visible and controllable. A political chatbot should display how many new factual claims it has introduced, how quickly it is producing them, and whether the user has asked for additional material. A citizen should be able to switch to a slow mode that presents one claim at a time, pauses until the evidence has been opened, and offers a neutral summary before moving on.
Sources should be attached to claims, not scattered as decorative links at the end of a long answer. The interface should separate established facts, contested interpretations, and predictions. It should also let the user request the strongest counterargument without forcing the person to leave the conversation and find an opposing system.
Political chatbots should identify the model version, the date of the governing instructions, and the organization responsible for the conversation. Campaigns, parties, advocacy groups and platforms should preserve sample transcripts for independent audits. Those audits should measure claim accuracy, correction behavior, source diversity, and whether the system changes its standards when arguing for different sides.
These measures are viewpoint-neutral. They do not decide which position is correct or prevent a campaign from making a forceful case. They address a procedural asymmetry: one participant can generate arguments at machine speed while the other remains bound by human reading, memory, and verification.
Disclosure alone will not solve that asymmetry. A label saying that the speaker is a bot tells the citizen who is talking, but not how the interface is shaping the exchange. A clearly identified system can still overwhelm the user with a sequence of claims that arrives too quickly to compare, challenge, or remember. Transparency must therefore cover the mechanics of persuasion, not merely the identity of the persuader.
A useful standard could be tested without deciding which political arguments are permissible. Does the system count and display new factual claims? Can the user stop automatic continuation? Are sources attached to the precise claims they support? Can an independent reviewer reproduce a sample conversation under the recorded model and instructions? These are measurable design questions, and campaigns could be compared on them before an election rather than after a controversy.
The limits of the research matter. The conversations lasted a median of about 14 minutes, participants were paid, and the tested political issues were British. Real-world influence will depend on exposure, trust, repetition, and whether people willingly sustain political conversations with a machine. That uncertainty is a reason to test safeguards, not a reason to ignore the mechanism.
Democratic deliberation requires more than access to an answer. It requires time to understand, compare, and contest the answer. We do not need to ban machine speech to protect that space. We need to stop treating the machine's ability to set the pace as if it were a neutral feature of the interface.
A democracy in which one side can generate an argument faster than the other side can read it is not necessarily better informed. It may simply be losing control of the clock.
Roney Lima do Nascimento is a mathematics educator, AI specialist and doctoral candidate in Pure Mathematics at the University of São Paulo. He writes about model evaluation, education, democracy and institutional capacity. His work has appeared in Folha de S.Paulo, Nexo Políticas Públicas, Congresso em Foco, EUobserver, GovInsider, HEPI and Educational Leadership.
The President Invited Iran to Bomb My Home