Greenhouse data from 2,832 job seekers shows candidates will accept AI interviews, but only with human review and clear rules
At least four in five job seekers in the United States, the United Kingdom, Ireland, Australia and Germany were rejected for a role in the past year without speaking to anyone at the hiring company, according to new research on AI in hiring from Greenhouse.
The Future of Hiring Candidate Report 2026 surveyed 2,832 active job seekers aged 18 to 64. Of the 302 Australian respondents, 85 per cent said they had been turned down without a single conversation. That compares with 81 per cent in the US and in the UK and Ireland, and 86 per cent in Germany.
Tom Hyde, head of sales and revenue for Asia-Pacific at Greenhouse told HRD the findings show both sides of the hiring market are using artificial intelligence (AI), and both are unhappy with the result.
“It’s the doom loop of candidates feel like they need to do it to stand out,” Hyde said. “Recruiters are drowning in volume because the candidates can use AI to apply so many times, multiple jobs, multiple roles, you know, create this volume.”
How AI in hiring created a volume problem
Candidate use of AI is now close to universal. According to the report, 84 per cent of Australian job seekers used AI in their search, alongside 87 per cent in Germany, 78 per cent in the UK and Ireland, and 77 per cent in the US.
Much of that use is aimed at beating automated screening. Among US respondents, 62 per cent tried at least one tactic to get past screening software. Of those, 40 per cent added skills or keywords they do not have, and a third inserted hidden text aimed at an AI reviewer.
Candidates are also working to disguise their use of the tools. Just over half (51 per cent) of Australian respondents rewrote an application so it would not appear AI-generated.
Hyde said traditional screening signals are losing their value as a result. He described a Melbourne-based customer who was recruiting a chief marketing officer. She had shortlisted 12 applicants, and six of them submitted cover letters.
“[They] started reading the cover letters, all six had at least one paragraph that was identical,” he said. “And this was for senior C-level roles."
Candidates open to AI interviews, with conditions
The research suggests job seekers are less hostile to AI interviews than many employers assume. Nearly half (48 per cent) had taken an AI-led interview in the past year. For each of the three formats (live AI conversation, written or chat questions, and AI-scored pre-recorded video), between 68 and 71 per cent of candidates who had experienced it said it let them demonstrate their skills well.
Human interviews are not faring much better. While 76 per cent said their most recent human interview let them show their skills, 90 per cent reported at least one problem. The most common, cited by 32 per cent, was the interviewer making up their mind on first impressions. The findings add to recent evidence on how a poor interview process is costing organisations talent.
Respondents were also asked what they would do if no human conversation were initially available:
- 36 per cent would prefer a written application only.
- 27 per cent would choose an AI-led interview.
- 10 per cent would withdraw.
Three factors were most likely to increase their willingness to take an AI-led interview: human review of responses (49 per cent), an explanation of how they would be evaluated (40 per cent) and the chance to ask questions (33 per cent).
Nearly all respondents (96 per cent) wished they could communicate something more clearly than an application allows. The most common was their ability to learn and adapt (52 per cent).
Hyde said candidates “want a human to actually review it”.
“You have to have that human decision-making process,” he said. “That’s the key part of it.”
That position aligns with Lendlease’s warning that talent teams can’t take the human out of hiring.
What should HR leaders do about AI interviews?
Candidates want rules. Between 88 and 92 per cent of respondents across markets want at least one legal requirement for employers using AI in interviews. The most popular were advance notice that AI will be used (55 per cent), human review before rejection (52 per cent) and an explanation of what the AI scores (48 per cent).
In Australia, private-sector employers still face no explicit requirement to tell candidates when an algorithm, rather than a person, is making decisions about their application, a gap raised during HRD’s roundtable with Sydney HR leaders on the ‘artificial applicant’.
Hyde’s advice to HR leaders is to start small. “Trial it for a month, two months. Trial it one or two roles and see if it fits for your business,” he said.
“You don’t have to go all in on it straight away. Pick a couple of like roles that you think would be a good fit. Get your business most importantly comfortable with the technology, and then you can roll it out and you’ll see a good response from the candidates.”
The cost of doing nothing is visible in the data. Among candidates who interviewed, 45 per cent said they had been ghosted by an employer.
“If you don’t hear back, it’s just super frustrating,” Hyde said.