LG CNS Adds a Generative AI Skills Test to Graduate Hiring

What happened

LG CNS, the IT services arm of Korea's LG Group, has opened its second-half graduate recruitment drive with a new hurdle in the written exam stage. Alongside the traditional aptitude and personality test that is standard in Korean corporate hiring, the company has introduced a "CNS-specific test" that measures two things: awareness of current technology and industry trends, and the candidate's actual ability to work with generative AI.

Crucially, this is not a one-size-fits-all exam. LG CNS will split the generative AI assessment into coding and planning tracks, depending on the role applied for. The implication is that developers will be evaluated on how they use AI while writing and revising code, while planning-track applicants will be judged on how they bring AI into problem definition and deliverable design. The company is hiring in the triple digits.

Why now

Over the past two to three years, generative AI has moved in IT services from "nice-to-have tool" to baseline productivity infrastructure. Code completion, first-draft documentation, and test case generation are increasingly assumed to involve AI assistance. Hiring processes, however, lagged behind. Writing "proficient with AI tools" on a résumé is a completely different thing from being able to design a prompt, evaluate the output, and catch where the model is wrong.

That gap is what this change targets: verifying a skill résumés can't prove. LG CNS framed the drive around securing people who can turn technology into real value in core businesses like AI and robotics. In other words, the company is not looking for candidates who find AI impressive, but for candidates who use it to raise the quality and speed of their output.

How this differs from standard Korean tech hiring

For years, graduate hiring at Korean IT and system integration firms followed a familiar template: aptitude test, algorithm coding test, then interviews. The algorithm test measures how accurately you can produce correct code under time pressure. It is a reasonable proxy for logical reasoning, but it has drifted further and further from how engineers actually work now — with AI assistants open in the next window.

The CNS-specific test is an attempt to close that gap, shifting the evaluation axis from "good without AI" to "good with AI." The hard part is grading. Separating the candidate's own judgment from the model's contribution in an AI-assisted deliverable is genuinely difficult, and how well that scoring design holds up will likely determine whether the new test is meaningful or cosmetic.

What it means for job seekers

First, there is one more thing to prepare. Grinding algorithm problems alone is no longer enough to clear the written stage. Practical experience — doing code review, requirement write-ups, or documentation alongside an AI tool — now doubles as exam prep.

Second, the explicit inclusion of "interest in current technology and industry trends" as a scored item is telling. It rewards candidates who actually follow and digest industry developments over those who simply stack credentials.

Third, choosing which role to apply for matters more. The split into coding and planning tracks means applicants effectively have to commit to a direction at the application stage.

Will other companies follow?

LG CNS is one of the larger graduate employers among Korea's conglomerate-affiliated IT services firms, and when a company of that size changes its screening format, rivals and the test-prep market tend to move with it. Coding tests themselves began as an experiment at a handful of firms before hardening into an industry standard; generative AI assessments could plausibly follow a similar path. At minimum, one thing is now settled: whether you can actually work with AI has become a scored item in entry-level hiring.