Colorado job hunters are using AI. So are employers, scammers and even the state’s labor department.
Matthew Boeckman, a hiring manager at one of Denver’s hottest artificial intelligence tech startups, shared a file on his screen where he’d pasted bullet points copied from seven resumes.
They look very similar, except for one which, he noted, “actually doesn’t have the line item that starts with the word Collaborated!” The other applicants used generative AI to create their resume. He knows this because a hiring manager asked Claude, the conversational AI chatbot from Anthropic, to create a resume from the job description. It’s one of the seven bullet points.
“They’re effectively identical in their phrasing, wording, cadence, tone, etc.,” said Boeckman, vice president of engineering talent at Magic School AI, which uses AI to help teachers with planning. “The other six are from real humans. Probably. Two of them are associated with LinkedIn profiles that have existed for less than a month.”
When a job applicant uses AI-generated content, it’s noticeable, he said, especially to an experienced hiring manager like himself. Of course, the use of AI isn’t necessarily frowned upon at a company with AI in its name. But doing so may not really help the job seeker, at least not at Magic School, which is currently getting an average of 900 applications per opening.
“You should use AI to help you with your resume. That’s reasonable coaching,” Boeckman said. “But sometimes the outcome is tough. One of these people might be brilliant for the role, but it’s a little tough to sort them from the other AI slop that we’re getting.”
Modern AI appears to be in wide use in the job market. A U.S. Chamber of Commerce post points to one study that found that 62% of job candidates use AI to apply for jobs, while another had about the same rate of hiring managers using it in the hiring process.
We’re in a Wild West-y time filled with unknown consequences. Bias from these LLMs, or large language models, trained on Wikipedia, social networks and who-knows-what else has folks inside and outside the tech industry pushing for more responsible AI. A new Colorado law goes into effect in June to address AI’s algorithmic discrimination. But the use of tools is growing fast, including here. According to Anthropic’s Economic Index, Colorado is its eighth biggest user of Claude AI nationwide.
AI definitely keeps popping up for Sherab Chödrön, who has been job hunting since moving to Denver in July. One experience involved an interview “with prepped questions, no person and a blank screen to talk into,” she said. More common? Rejections within 15 minutes of applying online. “No human could work that quickly at volume,” she said in an email.
She’s resigned to the trend that many other job seekers use it, as do hiring managers and employers.
“In addition, while I don’t use it this way, many folks are using imperfect AI tools to rewrite and create their materials to match job descriptions,” Chödrön added. “It’s yet another way to game the system, but it creates crap and noise along the way.”
So much more noise
Noise isn’t new, said Andrew Hudson, who back in the late 1990s started what has become Andrew Hudson’s Jobs List.
As online job boards like Monster.com and CareerBuilder were gaining momentum back then, job seekers could type in a job, location or salary and view multiple openings. But employers didn’t have the bandwidth to thumb through thousands of resumes. So applicant tracking systems became more prevalent to vet candidates by keywords and algorithms.
Clearly though, Hudson said, the tracking systems have become one of the biggest frustrations for job seekers. He knows plenty of folks who’ve put their “blood, sweat and tears into an application” and even tried to game it with keywords and phrases. “But you send it out and 10 minutes later, you get a note back saying, ‘Thanks, but no thanks.’ Clearly a human being has not read that resume or application. It’s spit out because you didn’t match certain criteria,” he said.
Today’s widespread use of AI-infused job boards makes it too easy for anyone to apply, especially the unqualified. LinkedIn’s job site is getting 9,500 applications per minute. We’re also in an economic environment where there are more unemployed workers than available jobs — and the latest U.S. figures don’t include furloughs or layoffs from the federal government shutdown, nor the recent cuts at Amazon, Meta and so many others.
In other words, it’s getting even harder to stand out. Hudson said he’s asked hiring managers about whether they end up missing out on a lot of great people.
“Their typical response is, ‘Yeah, but we’re getting the good enough,’” he said.
Magic School AI uses an applicant-tracking system to filter out the weakest candidates. Resumes without names. Messed up formatting. Job titles that don’t match the position they’ve applied for. No contact information.
“The way I would characterize it is that it’s a soft filter,” said Boeckman, who joined the company in June and has been in engineering for nearly 30 years. “We’re not rejecting anyone purely on the basis of what AI tells us in the applicant tracking system. But we’re very much trying to get out of 1,000, here are the 100 that we really need to look at with human eyes.”
With human eyes, they spot other types of unsavory candidates. Scammers and bots, and the overemployed — people who work at more than one company, often not well, and without disclosing it to their employers. Boeckman said he’s “personally found eight, maybe nine engineers in teams that I’ve led.”
Currently, he has one applicant that he suspects is double or triple dipping. The applicant’s resume says 12 to 15 years of experience but their LinkedIn profile was created a month ago.
AI is still allowed. Engineering candidates have a take-home coding exercise. They’re told they can use AI but they must be able to explain the work.
“Not frequently, but occasionally at that stage, we’ll see just crud and junk code that’s been generated by an LLM. Engineers talk about code smell. There’s a smell to that,” he said. “And you can see it. There’s repetitions of patterns used. There’s certain phrases and words that you’ll just see show up all over the place. We rarely reject completely based on that assessment but it’s a decision point for us.”
There’s an increasing backlash to virtual systems. Hiring managers have been left wondering if the person on the other side of the virtual interview is a deep fake. Or if the person has an AI chatbot on another screen spitting out responses. Cisco, McKinsey and Google are among the companies that have brought back face-to-face interviews with job candidates because of such concerns, the Wall Street Journal reported.
“After 25 years in hiring for startups and enterprise companies, what I see most clearly is that both hiring managers and candidates are treating AI like a cheat code — a shortcut to avoid the hard, deeply human work of connection,” said Dave Mayer, CEO of Technical Integrity, an executive search firm in Boulder. “The lesson from our perspective is simple: As an organization, if you care about building a thriving company culture and increasing engineering capacity, you need more humanity, not less.”
Hudson advises job seekers to use AI for feedback. Don’t ask it to write your resume. Rather, have it check your resume to see what’s missing. Modern AI searches can quickly pull up skills or experiences employers view as vital to a role. It at least adds another perspective.
“The power of AI is really the power of your creative prompts to help AI figure out exactly what you’re looking for,” Hudson said. “AI cannot replicate your authenticity. They can’t replicate your passion for something. They can’t tell the stories the way you can tell stories about your accomplishments and promotions and things that are really going to get people interested in your application.”
But, he admits, “it’s still a crapshoot,” akin to a human hiring manager having a bad day and overlooking a good candidate.
“I was talking to a job seeker the other day and they took the job posting they were going in for an interview for and asked AI, ‘Write me what are going to be the interview questions.’ They were able to use some predictive technology to help them prepare for the interview,” Hudson said. On another day, an employer told him, “‘We use AI to figure out what interview questions to ask.’ So, it’s like why do we even need human beings at all, right?”
AI used to find fits faster, interview and train job seekers
“Are you looking for full-time or part-time shifts?” a voice named “Eva” asks from the confines of a mobile app.
“Full time.”
“Got it,” Eva replies. “Which shifts are you interested in? We have morning, midday and evening shifts.”
So goes the sample conversation between a prospective job candidate and employer in a demo from Eightfold AI, a developer of “talent intelligence” that is making an artificially intelligent dent in how employers handle the growing number of job applicants, including in Colorado.
Eva is an agentic AI, a more autonomous system that can have a conversation with context instead of just spitting out a response to one question or prompt (that’s generative AI), like ChatGPT. The Colorado Department of Labor and Employment is considering using Eightfold’s interviewer to help unemployed workers improve their skills.
“We just had a demo of their AI interviewer because we’re interested in using that to help prep our job seekers who may not be so adept at interview skills,” said Mark Duey, strategic technology manager in the labor department’s Employment and Training Division. “It’s so job seekers can go through an interview for a specific job and get feedback from a case manager. It’s pretty powerful.”
AI interviewing systems are saving recruiters and candidates time. They can be scheduled 24/7 and hundreds can be done simultaneously, compared to humans having to schedule and conduct them. One nonprofit in British Columbia told The New York Times that it used Ribbon AI’s interviewer to screen 500 applicants for a fellowship program, compared to 150 the year before when human interviewers handled it.
Eightfold, a Silicon Valley AI company founded in 2016, is already a vendor of the state labor department. Its tech is used in the revamped job board Connecting Colorado, which anyone on unemployment must register for. Upload a resume or share your LinkedIn page, and the service will use AI to build a user profile.
And while a job seeker can search for specific roles, the system also looks for other opportunities the applicant may be qualified for based on 1.5 billion anonymized data sets Eightfold has collected over the years. It knows that folks with skills A and B tend to do well in jobs that require skill C, said Rebecca Warren, Eightfold’s senior director of talent-centered transformation.
“So while you may say that (you’re looking) for a software developer role, we may say, ‘Hey, you would be great as an implementation specialist. Or your background also fits with a role in presales,’” Warren said. “I have a background in talent acquisition but when you actually look at what I do, you see a whole bunch of different things. You’d see negotiation, communication and a bunch of other skills.”
Systems like Eightfold are also doing more than connecting dots between workers and jobs. They can be trained to spot inconsistencies in resumes, or use smart automation to speed up the hiring process.
After using Eightfold’s AI services to modernize its hiring system, power-management company Eaton said it saved $2.4 million. Half of the applicants for hourly positions also tended to drop out but are now sticking around because they receive immediate feedback.
Morgan Stanley automated 80% of its interview scheduling through Eightfold and saw a 50% reduction in the time to hire a new employee.
“Everyone comes to us with a little bit of a different story,” Warren said. “Bristol Myers Squibb used us to help develop skills so they understood who was in their system already that they should be moving internally. They now have three-quarters of their entire talent group engaged in some form of talent development.”
Colorado’s labor department is still considering whether to add Eightfold’s AI interviewing service. It would just be another job-training tool for an agency whose primary purpose is to help unemployed workers get hired.
“Once you get to extrapolate all of those skills, you’re recommended for different jobs,” said Duey, with the labor department. “It might say, you’re 80% of the way to this next job or career. … What it would do is open eyes and minds to like, ‘Wow, I might be eligible for a career path, instead of just popping around to some of the jobs that are low-wage or minimum wage.’”
AI in the job market is moving fast. The mid-2025 report from recruiter Career Group that found 62% of job candidates used AI to write cover letters, resumes or writing samples? The rate was 32% just six months earlier.
“It seems like we’re in the first or second inning of this,” said Evan Walden, who started his venture job-board tech site Getro in Colorado but now lives in New York. “The dream with AI is that it actually helps both sides find the right fit faster because people waste so much time looking for jobs and companies waste so much time recruiting. That is the ideal. I don’t think we’re there yet. It’s a tool. It’s like a hammer and it can be used for good and for bad.”