AI Hiring and the Rise of Job Portfolio Careers
AI hiring is weakening candidate trust. See why workers are adding income streams, what the survey data reveals, and how employers can respond.

Why Workers Are Building Job Portfolios in an AI Hiring Market
For a lot of professionals, a second job isn't just extra income anymore. It's a hedge against layoffs, inconsistent hiring, and a job search that can feel weirdly one-sided - send applications, then wait for silence.
In an Enhancv survey of 1,000 professionals with multiple full-time roles, 79.5% said a second paycheck protects them from unemployment and economic instability.
That response points to a bigger shift in the employment relationship. Candidates may run into automated screening they never asked to be part of, get rejection messages without any real human feedback, and see job listings that may not match actual openings. At the same time, some workers say AI tools have cut down the time it takes to do their work. If employers still reward hours more visibly than output, that extra capacity can quietly turn into another income stream.
The survey figures don't prove that AI hiring directly causes overemployment, and self-reported data has obvious limits. What the data does capture is how workers describe a market where trust is harder to maintain. This article looks at why employment is starting to feel more like a portfolio - what opaque hiring and "ghost" listings cost candidates - and what employers can change without giving up automation that actually helps.
Why are workers treating jobs like portfolios?
A portfolio approach to employment means relying on more than one source of work or income instead of betting everything on a single employer to carry you through the long term. That can look like two full-time jobs, a main job plus contract work, or a side business.
The logic is familiar: if one source fails, you're not immediately stuck. In financial terms, you'd call it risk management. In employment terms, it's a way to avoid being completely exposed.
In Enhancv's survey of 1,000 professionals with multiple full-time roles, 79.5% described a second paycheck as protection against layoffs and economic instability. About half said they would need a 21% to 50% pay increase to give up that second income. Another 12.9% said they wouldn't leave the second job no matter what pay they were offered. Those answers suggest this isn't always about maximizing salary. For some people, it's about reducing dependence on a single employer.
A response to uncertainty, not a universal career strategy
The data covers people who already hold multiple full-time jobs. It doesn't tell us how common that is across the broader workforce, and it doesn't confirm that every respondent had the same reason for doing it. Still, it challenges the idea that a second job is automatically just a short-term hustle or something confined to tech culture.
In the same survey, 50.4% of respondents said they worked mostly in the office and 26.5% said they worked hybrid. Education was the largest reported industry group at 16.6%, followed by operations at 12.9% and healthcare at 11%. Software engineers made up 2.3%. The takeaway: this isn't only about remote software work, even if the survey can't establish how representative each industry is.
The trade-off workers are making
Multiple jobs can provide income resilience, but they also come with real costs: schedule conflicts, fatigue, divided attention, and potential conflicts with employment agreements. Two roles can work for a while, but it may not stay sustainable forever. A second paycheck isn't "free insurance." It shifts more responsibility onto the worker.
For employers, the key isn't the claim that everyone is quietly overemployed. It's that workers may care more about income security than the traditional expectation of loyalty. When pay, location rules, or staffing plans can change with little warning, diversification starts to look less like greed and more like self-defense. A portfolio becomes more attractive as the risk of relying on one job feels higher.
How does AI hiring weaken candidate trust?
AI tools can help employers sort applications, schedule interviews, and manage large candidate pools. The trust problem starts when candidates can't tell whether a human or a system made the decision, what information was considered, or how to correct an error.
In a separate Enhancv survey of 1,066 U.S. job seekers, 50.5% said they received a rejection in the past year without human feedback. Of the 538 respondents who reported "silent rejections," 63.8% believed an algorithm made the decision. That's not proof that an algorithm was involved in every case - but it is the perception candidates report. And when applicants can't get an explanation, automated screening becomes the most obvious suspect.
Opacity turns efficiency into suspicion
The survey also found that 68.5% of candidates said employers rarely disclosed their use of AI. Meanwhile, 31.4% said they abandoned an application after a one-way video or chatbot screen. Nearly half, 47.7%, agreed that AI hiring tools are biased against their age, race, gender, or background. Candidates aren't passive in this either: 49.6% said they relied on AI during their job search.
That creates a lopsided process. Employers may use automation to assess applications, while candidates use automation to write or refine them. Both sides end up interacting through systems where the criteria aren't visible to the other. Automation may improve throughput, but throughput isn't the same thing as a fair or understandable process.
Human review should be a real control
A human in the loop only matters if the person can review the evidence, challenge an automated result, and actually change the outcome. A rubber-stamp review can be just another layer of process - without accountability.
Employers should be able to explain which stages use automation, what the tools are intended to do, and how applicants can request review or accommodation.
The reported age differences also deserve careful interpretation. Silent rejections were reported by 62.2% of applicants aged 25 to 34 and 61.5% of those aged 18 to 24, compared with 32% of those aged 55 and older. These figures point to different experiences in this survey; they don't, on their own, prove that age caused the difference or that an algorithm discriminated. They are still a reason to look closely at outcomes and at how candidates are communicated with - rather than assuming the process works the same way for everyone.
What do ghost listings cost job seekers?
A ghost listing is a job posting for a role an employer doesn't currently plan to fill. A listing can stay up because of outdated systems, pipeline planning, or internal uncertainty - so candidates can't confidently read it as a clear signal of intent.
But when employers advertise roles without a real opening, applicants pay the cost of the ambiguity.
In an Enhancv survey of another 1,000 U.S. professionals in March 2026, 51% of respondents with at least eight years of experience said they had encountered ghost listings. Across the survey, 37% reported direct out-of-pocket losses. Those costs can include travel, childcare, and other expenses. The bigger cost, though, is often unpaid time - tailoring applications, completing assessments, and preparing for interviews.
Work samples can become unpaid consulting
Among 411 senior respondents, 31% said employers had asked them to provide strategy decks or marketing intelligence before making an offer. A work sample can be a legitimate way to evaluate a candidate - especially for roles where judgment and execution matter.
The line gets crossed when the task looks like something the company could use as a deliverable, takes substantial time, and has no clear evaluation boundary.
Candidates also reported seeing a role reposted as new immediately after receiving a rejection email: 16.1% said this had happened to them. A repost doesn't automatically mean the original role was fake. It may have reopened, the criteria may have changed, or the listing may have been refreshed automatically. Without an explanation, though, the candidate is left to deal with the mismatch.
Put limits around the hiring process
Employers can reduce these costs with basic operational discipline. Keep postings current, clarify whether a role is approved and actively being filled, and close listings promptly when hiring pauses. For assessments, define the expected time, explain what will be evaluated, and avoid requesting work that could be used directly by the business.
Candidates can also set boundaries. Before taking on a large assignment, ask how long it should take, who will review it, and whether the company will use the submitted work. A reasonable employer should be able to answer. If the request looks like free consulting rather than a bounded assessment, it may be smarter to decline than to keep pursuing a role with unclear terms.
Why is overemployment not just a tech trend?
AI tools can make some tasks faster - but that doesn't automatically mean a worker has spare capacity. Workload depends on the job, the quality bar, review requirements, and how much coordination is involved.
In Enhancv's survey of multiple-job holders, 28% reported a significant reduction in workload from AI tools, while 8% said their workload had fallen by more than half. One in four said they met their primary targets in 30 hours or less.
These are self-reported estimates, not time-and-motion measurements. Still, they point to a mismatch many organizations haven't resolved: if a tool reduces the time needed to produce an acceptable result, does the employer reward the result - or keep treating desk hours as the main signal of value?
Output gains can change the employment bargain
Some respondents described practical ways they manage multiple roles. About 26.4% said they rely on speed alone, 15% used calendar blocking, and 14.5% carried separate devices. These figures describe reported tactics, not proof that juggling jobs is either easy or safe.
In many workplaces, holding parallel roles may violate policies, create confidentiality risks, or reduce work quality. Employees also have to manage what happens when deadlines collide.
Even with those caveats, there's a business implication. When AI reduces the time needed for certain tasks but compensation and expectations remain tied to hours, workers may use the difference to take on more work. That may not be what the employer intended as a productivity outcome. It can also reflect a system that measures presence more clearly than contribution.
Workplace rules can add to the pressure
Return-to-office mandates add another layer of friction. In a separate Enhancv survey of 1,000 U.S. employees, 72% viewed such rules as a way to encourage resignations. Another 36% said they had started a side hustle after their employer announced a return-to-work mandate.
These are perceptions and reported responses, not proof that every mandate is designed to force attrition. The useful takeaway is narrower: workplace policies get interpreted through whatever trust employees already feel. A rule imposed without a clear operational reason can look like the employer values control over outcomes. And when hiring feels opaque and job security feels fragile, income diversification starts to look rational.
What can employers do to rebuild trust?
The answer isn't to abandon AI hiring or promise permanent job security. Automated tools can reduce administrative work and help teams handle application volume. The question is whether the process stays explainable, reviewable, and proportional to the decision being made.
Make automation visible and accountable
Start by mapping where automated systems show up: application filtering, ranking, assessments, interview scheduling, or candidate communication. Tell applicants when automation plays a material role. Assign a person who can explain the process and review disputed outcomes. Keep records of what the system evaluates, and check whether results differ across relevant groups.
This isn't a guarantee the tool is unbiased. But it gives the organization a way to spot problems and respond. A vendor's assurance isn't a substitute for understanding how a system behaves in your own hiring pipeline, with your own role criteria and applicant pool.
Reduce candidate effort that doesn't improve decisions
Hiring teams should review each step and ask what decision it actually informs. If a task doesn't distinguish between candidates in a way the team can explain, remove it.
If a work sample is necessary, make it short, consistent, and clearly separate from work the company would otherwise pay someone to produce.
A practical process can include:
- Confirming that a role is approved and actively being recruited for before publishing it.
- Giving candidates a realistic timeline and updating them when it changes.
- Explaining whether AI is used and where a person reviews the result.
- Setting a time limit and evaluation criteria for assessments.
- Providing a channel for candidates to ask for accommodations or human review.
These steps don't require an expensive hiring platform. What they require is an owner for the process - and a willingness to measure candidate experience alongside speed and cost per hire.
Reward outcomes, not just visible activity
If AI changes how long work takes, managers may need to revisit how they define capacity and performance. That doesn't mean every efficiency gain should become reduced working hours. Some work still depends on collaboration, review, and availability.
But measuring hours by default can hide where tools actually improve delivery. It can also encourage employees to keep their capacity hidden instead of discussing it.
The employer's goal should be a sustainable agreement: clear expectations, fair compensation, and a realistic account of how work gets done. Workers may still want multiple income streams. The point isn't to shut that down - it's to avoid making distrust and financial exposure the main reasons people feel they have to.
Conclusion
Workers are building employment portfolios because a single job can feel less dependable - and because hiring processes often offer little explanation when an opportunity disappears. Enhancv's survey data reflects that feeling: second income is described as protection, candidates report silent rejections, and experienced professionals report running into ghost listings.
These findings don't prove that AI is the cause of overemployment, and they don't suggest every automated hiring process is unfair. What they do show is why opacity has consequences.
Employers can still use AI and also communicate clearly, keep job postings honest, limit unpaid candidate work, and reward outcomes - not visibility. Those are operational choices, not abstract promises. If workers are managing risk across a portfolio, employers should ask what in the employment relationship made that feel necessary.