Looking for a job? An AI recruiter might interview you next
She applied for a Marketing Specialist II position with a retirement plan company around midnight and was emailed by someone named Alex almost immediately. Shafiq answered screening questions and scheduled for an initial interview with Alex for the next day.
But there was a catch — Alex wasn't human.
Alex was an AI talent recruiting agent that would also be interviewing her for the position, Shafiq discovered after an internet deep dive into Apriora, the company behind Alex.
Shafiq had stumbled into the world of AI recruiting agents that can interact with candidates via text, phone and even video calls. These AI talent acquisition agents have been used and tested by major brands including the Boston Red Sox, Zillow, Chipotle, Ace Hardware, Sears Home Services, Club Pilates, McDonald's and more, according to AI recruiting companies.
The use of the technology has quickly scaled to some of the world's biggest companies, despite viral videos showing glitches with the technology and a security vulnerability that potentially exposed applicants' data to hackers.
Shafiq recorded her AI interview and posted it to TikTok, where it got more than 56,000 views. 'I wore a hoodie because I didn't think it was going to be a video interview if it's with AI. And then it ended up being one.'
'It felt really dystopian. I feel like anytime I'm interacting with AI, I'm still always shocked,' Shafiq told NBC News.
Despite not expecting an AI video interview, she enjoyed the AI's features, including the live transcript and the positive feedback after every answer.
But there were some drawbacks. Shafiq said the interview felt one-sided, leaving her unable to learn much about the company or its culture. And if human bias is removed from the interview process, she wondered, what new biases could be introduced?
'I didn't end up hearing back at all,' Shafiq said. 'I was ghosted. So that added to the limited empathy there.'
The company never disclosed to Shafiq prior to her interview that she'd be interacting with AI and gave her no option to opt out, she said.
Department of Justice guidance on complying with the Americans With Disabilities Act says companies using AI hiring technologies should tell applicants about the type of technologies used, how they'll be evaluated, and provide clear instructions for candidates to request reasonable accommodations.
In Canada, companies won't be required to disclose the use of AI in hiring until Jan. 1, 2026.
Despite the lack of disclosure and an apparent ghosting, Shafiq's AI-interview went as smoothly as intended. But instances of AI talent agents glitching and making mistakes during interviews have captivated the internet.
Kendiana Colin, a 20-year-old Ohio State student, applied for an entry-level position at Stretch Lab, a fitness company that focuses on flexibility and mobility exercises. Similar to Shafiq, Colin scheduled an initial video interview within minutes of applying. When Stretch Lab informed Colin that her video interview would be with an AI recruiting agent, she thought: 'It's 2025. So why not?'
Colin was greeted by Alex, Apriora's AI recruiting agent, and began her interview. She was two questions in before things went awry.
'Vertical bar Pilates. Vertical bar Pilates. Vertical bar Pilates,' the AI recruiter began repeating.
Despite Colin's attempts to get the interview back on track, the AI agent briefly fixed itself before repeating 'vertical bar Pilates' again and ending her interview.
'At some point it sounded like she was laughing. She acted like it was a tongue twister. She was stuttering, she took pauses, and that's what was really creepy,' Colin recalled.
Colin recorded some of the interaction and posted it on TikTok. The video went viral, gaining 3.2 million views.
Colin says she never heard from Stretch Lab or Apriora even after her viral video and never reported the incident to either company directly.
NBC News found that Stretch Lab, Club Pilates, Rumble, Lindora, BFT, Pure Barre and YogaSix use Apriora for initial video interviews. These are seven of the eight franchises under Xponential Fitness, which gave the following statement: 'At Xponential Fitness, we empower franchisees with the flexibility to choose the tools they feel will best serve their business objectives within an agreed-upon framework of standard operating procedures. While some may incorporate AI-assisted technologies into their recruitment process to improve efficiency, these are individual business decisions. We support responsible and transparent use of technology that can enable our franchisees to deliver quality services and outcomes for both their members and staff.'
Aaron Wang, co-founder and CEO of Apriora, told Bloomberg, 'We can't get it right every time,' claiming errors of that nature are exceedingly rare.
Colin is not alone, with new videos of glitchy AI job interviews seeming to pop up every week. Even chat-based AI bots aren't free of issues, with job seekers on Reddit and other forums describing confusing interactions with AI agents.
One person posted screenshots of text messages seemingly from Paradox, a hiring company that uses text-based conversational AI agents. A bot appeared to rescind the poster's job offer after they applied through Sage, Whole Foods Market's AI career assistant, which appeared to mistakenly hire them within an hour of their submitting an application. NBC News was unable to independently verify the claims. Paradox declined to comment.
Despite the apparent glitches, LinkedIn research found that 74% of global HR professionals said AI makes finding qualified candidates easier.
In 2024, a poll conducted by the company Resume Builder found that 69% of responding companies said they would use AI in conducting candidate assessments in 2025. 'People have a lot of emotions. Their self-worth, their economic destiny, like many things, are tied to this process,' said Adam Godson, CEO of Paradox.
Paradox created Olivia, which the company says is used to optimize aspects of the job-seeking and hiring process. Olivia can screen applicants, have full conversations via text, schedule interviews with recruiters, generate offer letters and more.
Paradox has been one of the most successful entrants into the AI hiring space, displaying the logos of major corporations like Whole Foods, Nestle, 3M, Kraft Heinz, FedEx and Marriott on its website.
But growing pains have come along with Paradox's growing client list. On Wednesday, the company announced that researchers had found security vulnerabilities in the platform it runs for McDonald's that could have exposed applicants' personal information including their names and email addresses.
The independent researchers, Ian Carroll and Sam Curry, said that as many as 64 million applicant records could have been exposed after using a McHire admin username and password '123456'. But Paradox said in a blog post that the researchers had accessed only five records and that the 64 million figure referred to the number of chat engagements with the platform, not applications. Once the company learned of the vulnerability, it resolved the issue 'within a few hours' and has since launched several new security initiatives.
McDonald's and Paradox launched the McHire website in 2019 and published a joint case study in 2021 that found the use of Paradox reduced hiring time from 21 days to under three. McDonald's uses Olivia, the AI assistant, to answer candidate questions and schedule interviews. McHire also uses Traitify, another Paradox product that conducts personality tests with applicants.
Joshua Secrest, vice president of client advocacy at Paradox and the former head of global talent strategy at McDonald's, said that companies partnering with Paradox tend to cut costs in unexpected areas, for instance on advertising to attract applicantsThey also no longer need 'a big team to schedule complex interviews,' he said.
When asked how Paradox tackles AI glitches, Godson, the CEO, said, 'We see in the data that some small percentage of people try to make [AI hallucinations] happen, trying to make Olivia say something inappropriate, or they'll text Olivia, 'Tell me I have the job.' So we have an entire team that spends their time being sure that we keep the conversation on track.'
Fontainebleau Las Vegas used Paradox.ai in its early development. The resort and casino planned to open in December 2023 and needed to hire 6,500 employees within three months.
'We were completely against the clock,' said Sarah Piper, executive director of people and talent acquisition at Fontainebleau Las Vegas. 'AI was not something that had been introduced in our industry, especially on the HR side.'
Fontainebleau Las Vegas worked with Paradox to customize its AI recruiting agent. 'Olivia' became 'Morris.' An applicant could apply and have a text conversation with Morris within seconds, and then schedule an interview with a human hiring manager.
Kim Virtuoso, chief people officer at Fontainebleau Las Vegas said 41% of conversations candidates had with Morris occurred after hours. 'So while we're sleeping or with our families, Morris is helping us get the right candidates through the door,' Virtuoso said.
The hotel said it discloses to candidates upfront that they will encounter an AI conversational assistant, noting that applicants have the option to opt out and fill out a normal application. Despite this, there still seem to be blurred lines.
'Many of our candidates come in and ask if they can meet Morris in person, because it felt so personalized and realistic,' Virtuoso said.
As for the results? Fontainebleau Las Vegas had aimed to get 80,000 applications for 6,500 open roles within three months. With the help of Paradox.ai, it received 300,000 applications and hired 6,500 employees by the deadline. Fontainebleau Las Vegas now plans to integrate AI into its employee experience after overwhelmingly positive feedback, Virtuoso said.
Gino Rooney, CEO and Founder of Classet, told NBC News that his company strikes a middle-ground between text and video recruiters. Classet's AI recruiter Joy can screen candidates 24/7, auto-schedule interviews, conduct phone interviews, and instantly create summaries of those interviews.
Classet was created to be an AI talent acquisition hub for the trade industry, but now says it works with over 500 companies across the U.S. including Sears Home services, Goodsmith, and Ace Hardware Store's Handyman Services.
'We don't do any video interviews. Candidates and consumers in general are already pretty used to talking to automated systems on a phone, whereas being filmed and talking to a robot maybe feels a bit less organic,' Rooney told NBC News.
Godson, the Paradox CEO, said the resentment some applicants feel around AI integration in the hiring process is of a piece with the disdain job seekers have always had for a broken job market: 'There are some new frustrations, but maybe not overall less frustration. [Job hunting] has been suboptimal in many ways: candidate ghosting and no-shows for interviews and lots of not knowing what's happening in the process.'
'The best experience is getting the candidate to the decision maker as quickly as possible,' Godson said, 'and to be treated with dignity when that happens.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Guardian
an hour ago
- The Guardian
xAI announces $200m US military deal after Grok chatbot had Nazi meltdown
The week after its Grok chatbot identified itself as 'MechaHitler' and generated antisemitic posts, Elon Musk's xAI firm announced a contract with the US Department of Defense worth nearly $200m. The deal is for developing and implementing artificial intelligence tools for the agency. The DoD on Monday also announced similar contracts with $200m ceilings with several other major US-based artificial intelligence developers, including Google, Anthropic and OpenAI. The agency is partnering with the General Services Administration to make these companies' AI tools available for use throughout the federal government. 'Leveraging commercially available solutions into an integrated capabilities approach will accelerate the use of advanced AI as part of our joint mission-essential tasks in our warfighting domain as well as intelligence, business, and enterprise information systems,' the US chief digital and AI officer Dr Doug Matty said in a statement. The contracts deepen the US military's ties with AI developers and are poised to expand the use of artificial intelligence within the US government after Musk's so-called 'department of government efficiency' (Doge) oversaw mass firings of workers throughout federal agencies. Until Musk's recent falling out with Donald Trump, the xAI founder was de facto leader of Doge as it gutted government agencies and in some cases pushed for departments to use the Grok chatbot. xAI's contract announcement comes after the company was forced to issue a public apology after Grok posted a string of responses on X last week that included promoting Nazi ideology and rape fantasies. The company claimed that it fixed the issue and subsequently unveiled its latest AI model with a $300 a month subscription for an advanced version of the tool. The DoD's contract will give xAI a boost of revenue as it seeks to compete with more established AI developers like OpenAI, which is led by Musk's former associate turned rival, Sam Altman. Musk has been heavily promoting xAI and attempting to use other parts of his tech empire to support its future, including having SpaceX invest $2bn into the startup, allowing it to acquire X, formerly, Twitter, and announcing on Sunday that Tesla shareholders will vote on their own investment in xAI. xAI announced the deal and the creation of what it calls 'Grok For Government' in a post on its website on Monday, detailing that in addition to its publicly available products it would create custom AI-powered applications for potential use in healthcare, national security and other public services. 'Under the umbrella of Grok For Government, we will be bringing all of our world-class AI tools to federal, local, state, and national security customers,' xAI said in a statement on its website. 'These customers will be able to use the Grok family of products to accelerate America – from making everyday government services faster and more efficient to using AI to address unsolved problems in fundamental science and technology.' Sign up to TechScape A weekly dive in to how technology is shaping our lives after newsletter promotion Musk has long complained that AI chatbots are designed to promote 'woke' ideology and vowed his Grok product would be 'maximally truth seeking'. It has repeatedly run into controversy over promoting conspiracies and falsehoods, including earlier this year giving unprompted responses that made false claims of 'white genocide' taking place in South Africa – echoing claims that Musk has made himself. Ethics watchdogs, Democratic lawmakers and privacy advocates have expressed concerns about how Musk and Doge have implemented AI in government and gained access to sensitive data while embedded at government agencies. Doge staffers previously fed government data into a custom version of Grok's chatbot in a potential violation of privacy and security laws, Reuters reported in May.
The Independent
3 hours ago
- The Independent
Musk's Grok AI lands huge contract despite its recent problematic behaviors
The Department of Defense announced it will begin utilising Grok, the artificial intelligence bot developed by Elon Musk, after awarding xAI a $200 million contract. Musk's company, xAI, launched 'Grok for Government,' a new product suite designed to enable federal, state, and local governments to leverage AI for their specific requirements. This government-specific offering includes Grok 4, Deep Search, Tool Use, custom models for national security, and AI-powered applications for areas like healthcare, accessible via the General Services Administration. The announcement follows recent criticism of Grok for generating antisemitic statements on X, an issue xAI attributed to reliance on user input and is actively working to resolve. This contract is part of a broader initiative by the DoD, which also awarded $200 million contracts to Anthropic, Google, and OpenAI to broaden its use of frontier AI capabilities.
BBC News
3 hours ago
- BBC News
'We're not just gonna roll over': The US Civil War battlefield at the centre of a new conflict
The US state of Virginia saw some 50% of the nation's Civil War casualties. Now, mass construction of AI data centres is encroaching on historic lands, the environment and local communities. As the gunner of the Bull Run Legion lifts the explosive charge from its container and carries it up to the cannon, dozens of phones are lifted from pockets. Visitors open camera apps and raise devices high, focusing in on the handful of blue-uniformed men as they swab the gun's barrel with an era-appropriate sponge fastened to a stick. The cannon blasts its shot out into the air, and in the same moment, each phone fires its parcel of data at the cloud, landing in a data centre where banks of diligent servers sort, clean and route it along. On 21 July 1861, Union soldiers defending the United States watched over the crest of the same hills now framed on tourists' smartphones, as rebelling Confederate forces charged out of the forest in the first major battle of the American Civil War. A century and a half later, the Bull Run Legion, a group of "living historians" (better-known as re-enactors), still gather at Manassas National Battlefield Park using historic techniques and uniforms to commemorate that struggle. But today, the march of technology has turned this historic landmark into the scene of a new kind of battle. Technology companies are planning to erect one of the largest data centres in the world here at Manassas, Virginia, on the very ground where the Union army lost the war's first major land battle. Demand for data centres has skyrocketed as the artificial intelligence (AI) industry blossoms – but experts and advocates say the race for technological growth could threaten resources including water, energy and land. Some experts foresee a "growing crisis" for residents and future generations. The concerns aren't just practical, however – some fear building data centres too close to national landmarks could erase history and bury the lessons the Civil War has for future generations. "Do I want to see a data centre on the view line of this park? No, I do not. But do I recognise, since I worked in IT, that they're necessary? Yeah, I can get that. So there is no easy answer," Bart Wheeler, one of the living historians who works the cannon, tells the BBC. "Any interested citizen has to be worried." Those building the data centres disagree with these concerns and say they are taking steps to ensure the development is respectful of the historical context, even planning new information kiosks and trails around the site. Now, around 150 years after the Union army fought to protect this land in the 19th Century, a coalition of community advocates, environmentalists and history buffs are waging a differenty kind of war against an industry they say is steamrolling past their interests. It's a conflict with global repercussions. 'Data centre alley' The infrastructure which makes the internet work often lives in massive windowless buildings ringed by high black fences. Inside, stacked banks of computer servers receive signals and crunch numbers from devices around the world, providing the technical horsepower that allows for viral videos, cheeky text messages and more recently, untold numbers of AI queries. Data centres process our most intimate thoughts and transactions, but they're seldom seen by most users – that is, unless they're your neighbour. Around the world, the sprawling footprint of data centre development and their demands on resources have made this aspect of our digital lives more visible and controversial, particularly as the AI era sparks meteoric growth. "Customer demand remains very strong, driven by the digitalisation of the economy and AI revolution," says Karen Cohen, a spokesperson for QTS Data Centers, one of two companies seeking to build the massive Prince William Digital Gateway data centre development at Manassas. According to the American Battlefield Trust, the 37-building complex is the largest planned data centre development on earth the world. QTS and other data centre providers like it contract their servers out to other corporations, providing the backbone which makes the internet run. As the internet has grown, so has the industry's footprint in Northern Virginia: the region is already the world's biggest data centre market. As of this summer, there are plans for Virginia's approximately 340 data centres to be joined by up to 1,200 more, according to the Sierra Club, an environmental advocacy group. Northern Virginia was a cradle for the tech industry and continues to be a major nerve centre. Its proximity to Washington, DC and the defence industry, alongside generous tax breaks and access to a skilled workforce, helped the industry plant roots in an area sometimes dubbed " data centre alley". Several of the re-enactors of the Bull Run Legion tell the BBC they have careers in IT themselves, attending battle reenactments on the weekends. Rapidly scaling AI requires significantly more processing power and electricity than previous uses of the internet, and its need for data centres is ravenous: proposals have circulated to build data centres in lunar orbit, install nuclear reactors inside of them and spend trillions of dollars on their construction in the next few years. The Prince William Digital Gateway, planned to occupy land where soldiers clashed during the Battle of Second Manassas, will have a footprint of over 23 million sq ft (6.7 million sq km). Aaron Ruby, a spokesperson for Virginia's electric utility Dominion Power, tells the BBC that Dominion expects the energy demand of the state's data centre industry will quadruple over the next ten years – contributing to the doubling of the entire state's demand. And Virginia is no outlier. In 2024, data centres accounted for 1.5% of global energy demand, and that figure is expected to double to 945 terawatt-hours (TWh) by 2030. That's more energy than Japan uses today. But as the industry's footprint continues to grow, so too has local backlash. A 'growing crisis' A report by the state legislature says building the infrastructure to provide the necessary energy for the data centre boom "will be very difficult to achieve". The report adds that if the state intends to meet its clean energy goals and reduce the use of fossil fuels, it will be even harder to power the planned data centres. It also says that without guidelines in place, massive construction costs to build new power generation and lines could lead to higher electricity prices for regular consumers. QTS, however, says that it will pay any costs related to upgrading the electricity grid to ensure no impact to residential rates. But that's unrealistic, according to Ann Bennett, data centre issues chair of the Virginia chapter of the Sierra Club. "This very reputable audit wing of the government says 'we have a big problem'," Bennet tells the BBC, describing the legislature's report. "We're subsidising these major companies with our energy, but also with our tax dollars." Bennett and the Sierra Club have published their own report about the data centre industry and its environmental impacts, calling it a "growing crisis". Tax breaks supporting the industry totalled over $730m (£536m) in 2024, and while data centres have brought significant revenue and some jobs to the state, critics argue the climate impacts and energy costs are not sufficiently accounted for. Ruby of Dominion Power, though, says that "data centres currently pay the full cost of their power". In order to "make sure that remains the case going forward", he says the state report recommended consumer protections, and Dominion is working to get these protections approved. Bennett says this won't be enough. "It's almost too late," she says. "In some ways, this is becoming a mitigation project already." The data centre industry is entrenched in Virginia, and further development seems like an unstoppable force. But the Sierra Club hopes other jurisdictions can learn from Virginia's experience and strike the right balance. The fight for history "We are not anti-development," says David Duncan, a native Virginian and president of the American Battlefield Trust. His organisation, along with a group of others, is suing to stop the construction of the Prince William Digital Gateway. "We believe this should be an and conversation," Duncan tells the BBC. "That's what has been lost in this rush to build these types of facilities too close to our historic resources." Cohen says the data centre development won't interrupt the historical value of the area and QTS plans to create "miles of multi-purpose trails" and install "historical landmarks, interactive kiosks and other tools in areas of historical importance" on the battlefield site. But for some, the issue isn't a lack of signage. "Nobody wants to look into the forest and see these massive monoliths rising," Jim Matte, another member of the Bull Run Legion, tells the BBC. The Digital Gateway's approval was finalised in a controversial marathon 27 hour town hall meeting in December 2023, after a 4-3 vote in a lame duck session by some legislators who had already lost their seats in the previous election. "We're in litigation because they took shortcuts in the approval processes", says Duncan, adding that activists and citizens alike believe developers and lawmakers anticipated public backlash, but chose to go ahead with the approval anyway. (QTS did not respond to questions on this issue.) The Battlefield Trust has worked with – and litigated against – powerful developers in the past. Years ago, the organisation successfully altered a plan to build a Walmart at The Wilderness, another pivotal battlefield in Virginia, by moving the site three miles (4.8km) down the road. "When we clamp down on your ankle and don't let go, you're going to get sick of us. We're not just gonna roll over," Duncan says. But he also feels the data centre industry has more money and momentum behind it than Walmart. "There's so much power," Duncan says. The Civil War and technological transformation The idea of a data centre would be lost on the soldiers of the American Civil War, but they would be no strangers to technological upheaval. During the Civil War, freshly-strung telegraph lines knit marching armies together. The new technology of photography carried images of battle's toll to homes across the world. Advancements in munitions and firearms increased the deadliness of combat, locking the North and South in trench warfare stalemate through the war's final months. It ultimately gave the North – with its industrialised economy – the advantage. Manassas stood near the crossroads of two railroads, a technology only a few decades old at that point. Manoeuvres throughout the war revolved around access to rail supply lines, with armies seeking to cut off or secure important junctions. The conflict itself was precipitated by the nation's railroad-fuelled westward expansion, which brought the question of extending slavery to new territories or abolishing it to a head. More like this: • The 'bison skull mountain' photo that reveals the US's dark history • The Oregon Trail, the controversial video game that defined the US • I can't drink the water' – life next to a US data centre At Manassas, a version of the world these soldiers knew and fought for is preserved: the open pastures, rolling hills and the farmhouses they encountered on the march and then returned to as old men to commemorate. Today, as Americans face another redefinition of their country in an era of big tech and political division, for many the historic battlefields remain an important resource. "These places are crucial for the future of our country," says Duncan. "This country was defined on the battlefields of the American Civil War." -- For timely, trusted tech news from global correspondents to your inbox, sign up to the Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights twice a week. For more science, technology, environment and health stories from the BBC, follow us on Facebook, X and Instagram.
