If measured purely by pitch accuracy, vocal consistency, and replicability, could artificial intelligence one day outperform a human singer? When that question was put to Im Seulong (임슬옹), a member of South Korean group 2AM, he neither dismissed the idea nor deflected with a simple declaration that human artists are irreplaceable. Instead, he offered a more considered answer: technically, AI will likely keep improving — but whether it can genuinely move people, stir emotions, and sustain long-term audiences is a different question entirely, and one that remains unresolved.
Im spoke with reporters on March 26 following his appearance at AI EXPO Taiwan 2026, where he addressed the future of entertainment technology. When asked who would win a head-to-head singing competition against an AI, his response was pragmatic. Technology and artistry, he said, are fundamentally different things. AI may well surpass humans on vocal control and technical precision, but entertainment is not ultimately about precision — it is about emotional impact. AI might be able to sing in tune, he said, but "whether it can make a person's heart beat, whether it can make someone smile" is a harder problem to solve.
AI Assists Creation, But Cannot Replace Emotion
"AI can assist in the creative process, but it cannot replace emotion itself," Im said. For an artist who has spent years performing on stage while also moving into the entertainment technology space, the statement reflected an industry read rather than a sentimental stance. AI will change how content is created and how performances are produced, he suggested, but it may not be able to fully replicate the uncertain yet genuine emotional exchange between a performer and an audience — at least not in the near term.
Im said he is not opposed to AI and has already begun integrating it into his work. He noted that the K-pop industry has started applying AI across multiple areas, particularly in image generation, video storyboarding, and pre-production simulation for music videos. For his own solo projects and future 2AM visual content, he said he is actively exploring how to use AI tools more efficiently, with a goal of releasing concrete results in the first half of the year.
He elaborated that AI's most practical contribution to music production currently lies in the early testing and verification stages. From songwriting and composition to arrangement, instrumentation, vocal layering, and overall atmosphere simulation, AI can rapidly generate near-finished test versions before a creator commits to a final direction. This does not mean AI replaces the creator, he said — rather, it allows creators to identify possibilities faster, eliminate directions that do not work, and save time and cost. The critical decisions, he emphasized, still come from human experience and emotional judgment.
AI Cannot Replace Live Artists — But Can Extend Their Reach
What makes Im's perspective noteworthy, however, is not how much he embraces AI, but where he draws the line. Reporters at the event asked about the trajectory of AI-driven entertainment in light of a clear post-pandemic trend: while online concerts, extended reality experiences, and digital avatars expanded rapidly during COVID-19 lockdowns, audiences returned to live events once restrictions lifted. Fans still want to see real artists in person, and many have expressed a preference for receiving messages from the actual person rather than an AI stand-in. Given that reality, how can AI-powered entertainment be sustained?
Im's answer was direct: face-to-face interaction remains deeply important. For him, the shared energy, physical presence, and emotional flow between an artist and an audience in the same space represents one of entertainment's core values. Live performances are not relics of a previous era, and they will not lose meaning simply because AI exists. If anything, he suggested, rapid advances in digital technology make the irreplaceability of live experience more apparent, not less.
That said, he does not see this as evidence that AI has no role. He defined AI's function as one of expansion rather than replacement. Artists cannot be available around the clock, and geography and time zones create natural limits on reach. AI and digital content, he argued, can fill in where physical presence is impossible. The future of entertainment, in his view, is not a choice between real artists and AI — it is more likely that real artists remain at the core while AI handles expanded peripheral interaction.
Artist IP and Digital Identity Need Clearer Rules
That framing transforms the question of whether AI can outperform a human singer from a technical competition into a more complex question about industry structure and emotional authenticity. Measured purely by voice replication and algorithmic precision, AI will continue to improve. But audiences love a song, remember a performance, or follow an artist over years for reasons that rarely come down to technical perfection — often it is because the work connected with a particular emotion or memory at a specific moment.
Im also raised what he described as an increasingly unavoidable question: how an artist's voice, likeness, and digital identity should be used in the future. His position was cautious. The issue, he said, cannot be addressed in a sentence or two, because an artist's intellectual property is not simply material that can be cleanly segmented and licensed — it is a subject built on emotional relationships. Future frameworks for managing these questions must address creator rights, user rights, and the ethical boundaries on the consumer side with considerably more specificity.
That caution makes his overall position more coherent. Im is not arguing from uncritical technological optimism, framing AI as an all-purpose solution. Nor is he arguing from complete resistance, treating all digitization as a threat. His argument is that AI has already entered the entertainment industry and will reshape creation, production, and fan interaction at increasing depth — but that the foundational emotional connection between people is the element that cannot be removed from the equation.
He Asks ChatGPT His Daily Fortune — in a Colder Tone
On a more personal note, Im disclosed that he occasionally encounters factual errors when using ChatGPT, but finds those mistakes less significant than the genuine convenience AI provides in daily life. He mentioned that he has asked ChatGPT in the morning whether he would feel nervous that day, and received a response suggesting that his extensive stage experience meant he would likely be fine. He described the exchange as amusing.
He added that he sometimes asks AI about his daily fortune, and has gone so far as to adjust ChatGPT's response style — requesting that it reply in a calmer, more realistic, and even somewhat blunt tone. These seemingly minor personal details point to something broader: AI is not only entering creators' workflows, but also their emotional processing and decision-making rhythms.
Im also noted that AI may open new opportunities for small and mid-sized entertainment companies. Music video production, visual planning, and content creation have historically required significant shooting and post-production budgets. If AI can effectively lower the barrier to entry in image generation, visual effects simulation, and pre-production testing, smaller teams would be able to concentrate more resources on story, emotion, and core content rather than being constrained by high production costs.
On the original question — can AI outperform a human singer — Im's answer, across the full course of the interview, was clear: AI will keep improving vocally and in content production. But entertainment is not a technical exhibition. What makes someone remember a song, return to a performance, or follow an artist over years has never been accuracy alone. It has been whether the emotion genuinely arrives. Technology can keep closing the gap, and may surpass human performance on certain measurable dimensions. But singing into someone's heart, at least for now, is not a problem that model parameters alone can solve.
Social Media Links & Editor Tag - Penny
