As generative artificial intelligence becomes a main counselling channel for job seekers, concern is growing that companies with low online visibility could be excluded entirely at the hiring stage.

On June 22, local time, IT media outlet ITmedia reported that some companies confirmed their names did not appear at all when they asked large language models (LLMs) to recommend companies in their industry.

Behind the growing concern in hiring is an actual case of fewer applications. One company accepted applications for summer internships for students expected to graduate in 2028 after Golden Week, but the number of applicants stayed at about half the usual level. A subsequent internal review found that about half of respondents to the previous year’s new-hire interview survey had used LLMs in their job search activities.

When the company asked multiple LLMs about its industry and recommended companies, its name was not mentioned at all. The company saw reduced web exposure from cutbacks in promotional spending as one of the causes.

At the same time, LLMO is emerging as a new task in recruitment. LLMO refers to optimisation work aimed at getting generative AI such as ChatGPT, Gemini and Perplexity to cite or recommend company information. In the field, awareness is spreading that "LLMO, once discussed only in marketing, is starting to affect recruitment, too."

Students’ use of AI is already hard to ignore. A survey by Japanese job-support company Mynavi found that 2 out of 3 job seekers expected to graduate in 2026 said they use AI in their job search. A survey by employment services company Gakujo also put the share of students who have used generative AI during job preparation at 40 percent. Another survey found that about 70 percent of Generation Z do not use generative AI in job-hunting and job-changing activities, showing a transitional phase in which usage is expanding while a non-user group also persists.

Even so, a structural problem lies behind the rapid spread of LLMs: job seekers lack reliable counselling channels. The capabilities and operating standards of university career support centres vary widely by school. Some universities enjoy strong student trust, while others struggle to assign staff as demand concentrates on cover letters and interview preparation. At some universities, students learn about related organisations only just before graduation, and some have no employment department at all.

There is also criticism that the private job-counselling market is not a stable alternative for students. Talent introduction firms for entry-level hiring are divided between those with dedicated databases and those with entry-level databases for first-time workers. Even under the same company name, the former may show technical jobs first while the latter may present postings such as mobile phone sales jobs first. That means a student’s career starting point can change depending on which banner the student clicks.

Some also say that even firms specialising in entry-level hiring have increasingly steered candidates toward more profitable areas. Around 2024 to 2025, cases stood out in which applicants were funnelled to consulting firms or some system integration (SI) companies with high referral fees and easier hiring outcomes, or were actively connected to SES and staffing firms that finish selection in a single day. Many social media (SNS) accounts promoting 'career counselling' are in practice closer to affiliate revenue purposes, and in many cases their actual work experience is unclear.

In this situation, LLMs have become counselling counterparts for students that they can call on anytime and that appear neutral on the surface. For companies, the issue is not simply increasing hiring promotion. They also need to see that the first information channels students encounter are shifting from search and job fairs to AI answers. That is increasing the likelihood that companies with a weaker online presence will face greater disadvantages in future hiring competition.