Workforce conversations are moving past the familiar tropes of whether artificial intelligence will simply eliminate jobs. A more immediate shift is already taking shape. AI is changing what people need to know before they are hired, what employers expect them to do once they arrive and how much information workers can have about the terms governing their employment.

That theme surfaced repeatedly at Ai4 2026 in workforce programming and in two sessions I moderated, one with Lift Our Voices cofounders Gretchen Carlson and Julie Roginsky and Brigham Young University professor James Gaskin on using AI to help workers understand employment agreements, and another with New York Jobs CEO Council Executive Director Kiersten Barnet on talent and opportunity.

The bigger question is no longer just what AI does to a job. It is what AI does to a worker’s position throughout the employment cycle. Does it make it easier to enter the workforce or raise the bar? Does it give employees stronger skills or simply ask them to produce more? And when technology lowers the cost of expertise, does that new knowledge flow mainly to employers, or can workers use it to gain more control over their careers and working conditions?

Before The Job, AI Literacy Is Becoming Part Of Employability

The first disruption happens before somebody receives a paycheck. The impacts of AI are reaching far beyond people working in technology fields. A marketing graduate may now enter an interview expected to explain how they use AI. A midcareer employee may need to learn how to supervise AI generated work rather than produce every first draft personally. A new hire might use AI to inspect a contract containing forced arbitration or an unusually broad confidentiality clause before signing it.

Barnet leads the New York Jobs CEO Council, an organization founded and funded by 25 major private sector employers in New York City. Its members operate in financial services, healthcare, technology, insurance, professional services and other sectors. She said member companies have hired more than 65,000 low income New Yorkers during the past five years. That gives the Council an unusually direct view into what large employers are actually hiring for.

Barnet described AI literacy as becoming “table stakes,” including for workers who never intend to become engineers or data scientists. A marketing major, sociology student or finance graduate needs some feel for what these systems can do, where they fail and how they fit into actual work.

The numbers Barnet shared challenge a popular assumption about entry level work. Member companies hired about 17,500 entry level workers last year, she said, roughly level with and slightly above the prior year. Entry level roles have increased as a share of hiring over the past five years.

“If you ask any of our CEOs, they will tell you they don’t know what it’s going to look like in the future,” Barnet said. “That’s, I think, an honest answer anyone can give you, but that it will change.”

There is an awkward problem in getting those entry level workers ready for the AI-enhanced workplace. Many colleges are still deciding what acceptable AI use looks like in the classroom. Students may hear warnings about cheating or prohibited tools. Then they interview with an employer interested in the opposite question.

“We hear from our employers that they want to hear from students or entry level roles, how are you using AI? How do you use it in your daily life? How are you playing with it? How are you building with it?” Barnet said.

For a new worker, those signals can be maddening. Do not use it here. Be ready to use it there.

That disconnect could become one of the harder workforce problems created by AI. Employers can revise job requirements quickly. Academic programs tend to move on a much slower clock. Barnet’s work with New York’s public colleges is partly an attempt to close that gap by giving educators a clearer signal about which skills companies want.

AI literacy is only one part of what employers are asking for. Barnet said recruiters continue to push for skills that are harder to automate and harder to teach in a conventional classroom.

“The skills that they are really hungry for are things like critical thinking, work ethic and problem solving and collaboration, and public speaking and things like that,” she said. “Those are harder to teach in the classroom, at least the way we are used to teaching.”

Barnet offered an intriguing example of what that could mean for recruiting. In one recent internship cohort, about 1,000 CUNY students participated among roughly 10,000 interns at Council member companies. She said those CUNY students were 22% more likely to receive full time offers at the end of the summer.

Barnet was careful to separate the data from her interpretation. Her theory was that many CUNY students arrive with experiences conventional recruiting filters often miss.

“They’ve all had time management skills. They’ve had a job. They’ve had a boss. They may have been a boss. They’ve been yelled at by a Starbucks customer,” she said, pointing to customer service, responsibility and real world experience as skills employers value.

The problem is who gets access in practice.

“Transformations favor the skilled,” Barnet said. She argued that AI can help spread access to skills, but only if schools, colleges and training systems make it part of the normal education pipeline.

“If it is not baked into whatever our system for education is, K through 12, higher ed, trades, whatever that might be, it will then always need to be something that people are accessing outside of the system, which generally then will only widen opportunity gaps,” she said.

That reframes the workforce risk. The use of AI, or lack thereof, can create a skills gap before two candidates ever submit an application. Students with better devices, more time, stronger professional networks and knowledgeable parents or mentors can get good at these systems earlier. Others may technically have access to the same tools and still enter the labor market several steps behind.

After The Offer, AI Can Change Who Understands The Fine Print

While on the one hand AI can shift power to the employer, in other ways, sessions showed that AI can also help shift power back to the employee. This is especially the case when it comes to employment contracts.

Carlson and Roginsky, both veteran journalists and television commentators, founded Lift Our Voices after experiencing firsthand how workplace silencing mechanisms can shape harassment cases. Carlson, a former Fox News anchor, sued then Fox News chairman and CEO Roger Ailes in 2016, alleging sexual harassment and retaliation after she rejected his advances. Her case helped trigger a broader reckoning inside the network, the growth of the #MeToo movement and preceded Ailes’ resignation.

Roginsky, then a Fox News contributor and political commentator, filed her own lawsuit in 2017 alleging that Ailes sexually harassed her and retaliated against her after she rebuffed him. Both women later became outspoken critics of forced arbitration and nondisclosure agreements, which they argued can prevent workers from publicly challenging misconduct. Their advocacy through Lift Our Voices helped build momentum for legislation, the Ending Forced Arbitration of Sexual Assault and Sexual Harassment Act and the Speak Out Act, both enacted in 2022.

Prospective employees routinely receive contracts and workplace documents containing language they may not fully understand. Forced arbitration, confidentiality provisions, nondisclosure agreements and non disparagement clauses can carry consequences that are difficult to appreciate when somebody is excited about a new job.

Carlson made the problem personal in the session.

“I didn’t understand what I signed at Fox News,” she said, explaining that the forced arbitration provision in her employment agreement could have kept her case outside public court proceedings.

Her point was broader than her own experience. Carlson said the organization encounters the same lack of awareness from senior executives through minimum wage workers.

Lift Our Voices built its recently released, free LOV Where You Work AI Tool around that information gap. Workers can upload employment related documents and identify language tied to forced arbitration, NDAs and other forms of workplace silencing.

This is where AI’s role in employment becomes more interesting. Employers have obvious reasons to use AI for recruiting, productivity and workforce management. Employees have reasons to use it too.

A worker who cannot afford to have an attorney read every employment document may still be able to spot a clause worth questioning. Roginsky said the tool can give people suggestions about what they might do after discovering such language, whether they find it before signing or later.

The tool does not erase the imbalance between an individual applicant and a large employer, but it can change the amount of information each side brings to the table.

The Shifting AI Employment Conversation

Ai4’s workforce discussions painted a more complicated picture than the familiar jobs versus automation debate. AI is beginning to shape the employment relationship at every stage, from the skills candidates need to get hired, to the judgment workers must exercise on the job, to the information they can use when evaluating the terms of employment itself.

That changes the question. The next phase of the workforce debate will be defined not only by how many tasks AI can automate or how many jobs it may displace, but also whether workers can use these systems to gain skills, make better decisions, understand their rights and move into stronger positions as work changes around them.

If the sessions and conversations at Ai4 were any indication, that shift is only starting. The most consequential measure of AI’s impact on employment may be how much more informed, adaptable and capable workers become in response to smarter machines increasingly making their impact on employment.