What If DEUX's Voice Came Back?
Imagine a voice crossing decades and returning.
DEUX was a South Korean pop-hip hop duo that debuted in 1993 and disbanded after their third album in 1995.
In 2025, Lee Hyun-do (the surviving DEUX member and producer) announced a plan to use AI to recreate the late Kim Sung-jae’s voice and release a “DEUX 4” album.
Immediately after the announcement in August 2025, the news spread online and sparked a fierce debate.
However, the atmosphere around the project felt heavy.
Organizers say they discussed the plan with Kim Sung-jae’s family before proceeding.
Still, many fans feel confused, and some have protested.
The idea that AI voice restoration has moved from science fiction to reality has made emotions more complicated.
From a Small Stir to a Big Wave
The project began as technical curiosity, but the ripple became a wave.
Over the past three years, the amount of speech data available and the quality of synthesized voices have improved enough for real-world use.
The DEUX 4 project plans to extract Kim Sung-jae’s vocal traces from existing recordings, train AI models on the remaining samples, and generate new lead vocal tracks.
If successful, a commemorative release timed for the 30th anniversary in 2025 would be a major cultural moment.
Moreover, the technical effort aims to go beyond simple mimicry.
In music, AI voice restoration can combine with composition, arranging, and production to create new modes of creativity.
Lee Hyun‑do said in interviews he intends to “respect Kim Sung‑jae’s timbre and phrasing as much as possible” when working.
That remark suggests the team sees consent from the family and input from experts as essential to the project.
A City Choosing a New Challenge
The controversy runs deeper than it looks, so debate continues.
This project sits at the intersection of technology, culture, law, and ethics.
Proponents argue AI voice restoration can preserve and extend an artist’s legacy.
On the other hand, critics warn it risks violating posthumous rights and personal dignity when the deceased cannot consent.
Meanwhile, commercial concerns loom large.
When an AI-recreated voice is sold in the music marketplace, critics ask where commemoration ends and commercialization begins.
Copyright, publicity rights (the right to control one’s likeness), and who owns AI-generated material remain legally unsettled.
Lawyers say outcomes will likely be decided case by case.
Why Do People Split Into Two Camps?
The dispute digs into deep feelings, so the argument lasts.
Supporters view the process as technological progress and cultural preservation.
They say AI restoration is a new way to pass music to present and future listeners.
When projects proceed with family collaboration, supporters see that as respectful and modern—an alternative to traditional memorials.
Also, proponents point to empirical gains.
Current AI voice synthesis can learn a singer’s tone, articulation, and breathing patterns from archival audio.
Advocates say samples can be verified against concert recordings and studio masters to ensure musical quality.
Under this view, a DEUX 4 album could bridge past and present with fresh arrangements.
On the other hand, opponents raise basic ethical objections.
Recreating a voice when the person cannot give consent feels like an intrusion to many.
Some fans argue that artificial reproduction replaces the person rather than honoring them.
They prefer noncommercial tributes—live memorials, fan-led playlists, or documentary projects—to recreated recordings.
Opponents also link their view to historical precedents.
Art restoration debates have long pitted authenticity against reconstruction, and music has faced copy and forgery controversies too.
Thus, critics worry AI vocal reconstruction blurs the line between artistic restoration and commercial replication.
They insist public consensus and legal frameworks are necessary before such practices spread.
The Pro Side: Preservation and Possibility
In brief, this is where technology meets remembrance, and expectations are high.
Supporters emphasize technical feasibility.
They note that modern AI can reproduce not only a voice’s timbre but also subtle delivery habits.
Lee Hyun‑do and the production team say they collected dozens or even hundreds of original recordings to build a reliable dataset.
They plan to have the results reviewed by family members and vocal experts before release.
Furthermore, cultural preservation matters.
DEUX helped shape 1990s Korean pop and hip hop, and Kim Sung‑jae left a powerful musical imprint in a short career.
AI restoration could adapt that heritage to today’s listening platforms, where a new single or album still drives discovery.
From this angle, DEUX 4 might spark renewed interest and cultural rediscovery.
Finally, supporters offer a business rationale.
The music industry needs sustainable ways to manage legacy catalogs.
With clear contracts protecting family rights, transparent revenue sharing, and limits on use, AI restoration could be one model.
Proponents argue that if norms and safeguards are in place, the approach can become an ethically grounded option.
The Con Side: Consent, Authenticity, and Risk
Yet opposition draws on stronger emotions, so it is not easily quelled.
First, the absence of explicit consent is the core ethical worry.
We cannot know whether the deceased would have approved a synthetic recreation.
Even when families consent, social consensus can remain fragile and fandom conflict may continue.
Second, technical limits raise authenticity concerns.
An AI-generated lead vocal might mimic the original, but it can miss the tiny expressive cues that give a live performance its emotional charge.
A produced track could therefore fail to convey the original feeling or, worse, distort memories of the songs.
Third, legal and social precedent worries persist.
If AI vocal restoration is permitted here, others might use it without permission, setting a risky precedent.
Opponents argue that weak rules could erode cultural dignity and public trust.
They call for stricter laws, clearer rights, and broader public discussion before such practices become routine.
In short, critics combine moral appeals and institutional safeguards in a call to slow down and legislate cautiously.
Deep Dive: Causes, Reactions, and Paths Forward
We need to see the issue in context, so the analysis must be layered.
First, causes: a mix of emotional longing and technical possibility.
Kim Sung‑jae’s sudden death in 1995 left fans wanting more, and decades of admiration built unmet demand.
In 2025, AI offered a way to respond to that desire.
Second, internet reactions are multi-layered.
Supporters celebrate a new listening experience and an expanded memorial practice.
Detractors emphasize ethics and authenticity, and social media amplifies both sides.
Neutral observers predict the controversy will push law and policy debates forward.
Third, the project’s future depends on norms.
To set a positive precedent, the team would need explicit family consent, transparent contracts, independent verification, and published ethical guidelines.
If those conditions are met, AI restoration could join cultural preservation tools.
If not, controversy may linger and legal disputes could follow.
Conclusion: A New Form of Commemoration—If Done Right
The core is agreement and transparency, so careful process matters.
AI voice restoration can deliver both technical achievement and cultural value.
Yet the lack of posthumous consent and the need for emotional respect remain unresolved.
If Lee Hyun‑do and the production team honestly incorporate the family’s input, invite external oversight, and publish clear rules, this project could become a model.
On the other hand, procedural flaws or overt commercial focus could deepen fandom divisions and spark legal fights.
Ultimately, society must balance technological progress with ethical responsibility.
The DEUX 4 project might be more than an album release; it could be an experiment in how we honor artists in the AI age.
Do you think that recreating a singer’s voice after their death can be justified as a form of commemoration?
Using AI to recreate Kim Sung‑jae’s voice for a DEUX 4 album is both an opportunity to reframe a 1990s musical legacy and a challenge to ethics and rights. Clear family consent, transparent procedures, third‑party review, and published ethical standards are essential. With those safeguards, the method could become a new form of cultural remembrance; without them, controversy and legal uncertainty will remain.
