Specialty Digest

Global English editionLearners’ Study Dashboard — in development
The Knowledge Library for All

Digital Human

nounGenerative AIalso virtual human, AI human, synthetic human

In one line

A digital human is a highly realistic computer-generated person, often powered by artificial intelligence, that can speak, express emotion and interact with people.

A digital human is a highly realistic computer-generated person that can speak, show emotion and interact with people, often powered by artificial intelligence. It goes beyond a simple AI avatar by aiming for lifelike skin, eyes, hair, voice and body language, and by combining these with a conversational brain such as a large language model. Digital humans appear as virtual customer service agents, brand ambassadors, teachers, health coaches, news presenters and characters in games and films. Companies sometimes call them virtual humans, AI humans or digital people.

History of digital humans

Computer scientists began modeling human faces and bodies in the 1970s and 1980s, with early work such as Frederic Parke’s animated face in 1972. In 1985, the animated short Dire Straits and the character Max Headroom brought attention to synthetic personalities. Realistic digital actors appeared in film, from the 2001 feature Final Fantasy: The Spirits Within to Gollum in The Lord of the Rings and the young faces of actors created by performance capture and visual effects. Researchers in virtual human systems developed embodied conversational agents in the 1990s and 2000s. In the 2010s and 2020s, real-time engines such as Unreal Engine with MetaHuman Creator made detailed human models available to small teams, and neural rendering, generative adversarial networks and diffusion models allowed lifelike faces from photographs. Language models gave digital humans the ability to converse freely.

How digital humans are built

  • Modeling the body and face: artists or scanners capture a person’s shape, skin texture and hair, or tools generate a character from parameters. Photogrammetry and 3D scanning record detail from many photographs.
  • Rigging and animation: a digital skeleton and facial controls let the character move. Motion capture records actors, while AI predicts gestures and expressions from speech.
  • Rendering: real-time engines draw skin with subsurface scattering, realistic eyes and hair at interactive speeds. Neural rendering can synthesize faces directly from learned models.
  • Voice: text-to-speech or a cloned voice speaks the lines, and lip-sync models align the mouth with the sounds.
  • Intelligence: speech-to-text hears the user, a language model plans a reply, a knowledge base supplies facts and an emotion model chooses expression and tone.
  • Delivery: the digital human runs on a screen, kiosk, website, app, mixed reality headset or video stream.

Types of digital humans

  • Scripted digital humans: deliver prepared content, such as news or training.
  • Interactive digital humans: converse in real time using AI.
  • Digital doubles and twins: replicas of real people, for example actors for film or experts who want to scale their time, created with permission.
  • Fictional digital characters: original virtual people such as brand mascots and virtual idols.
  • Embodied agents in virtual and augmented reality: characters that share a virtual space with the user.

Applications

Retailers, banks and telecom firms use digital humans as friendly guides in apps and in stores. Healthcare projects test digital humans as patient coaches and as simulated patients so medical students can practice interviewing. Education providers use them as language partners and tutors. Media companies use virtual presenters and virtual idols, particularly in East Asia. Studios use digital humans to de-age actors and to complete scenes. Museums create historical figures that answer questions. In training for emergencies or difficult conversations, they offer a safe place to practice.

Benefits

Digital humans can provide personal, human-like service at any hour, in many languages, and at scale. They can make complex information friendlier and give people who prefer face-to-face interaction a more natural interface than a text box. They also allow practice of rare or sensitive situations without risk to real people.

The uncanny valley and trust

The uncanny valley, an idea proposed by roboticist Masahiro Mori in 1970, says that as a figure becomes more human-like, people respond more warmly up to a point, after which near-perfect but slightly wrong features cause unease. Designers deal with this by choosing a stylized look, polishing eye movement and timing, and by being open about what the character is. Studies of trust show that realism is less important than competence, consistency and honesty about limits.

Ethical and legal issues

  • Consent and rights of likeness: digital replicas of real people need their permission, and many countries are strengthening rights over voice and image. Actor unions have negotiated protections for digital replicas.
  • Deception: people should know when they are interacting with a synthetic character. See deepfake.
  • Emotional influence: lifelike characters that seem caring can influence decisions and create attachment, which calls for caution in sales, politics and companionship.
  • Bias and representation: the range of available faces, skin tones, accents and ages affects who feels represented.
  • Data protection: faces, voices and conversations are sensitive data.
  • Afterlife and grief: recreating deceased people raises questions about dignity and the wishes of families.

A worked example

A bank wants a digital human to help new customers set up accounts on its app. The team designs a friendly, clearly stylized character, connects it to a language model restricted to approved product information and adds a handoff button to a human adviser. They test it with users of different ages and accents, label it as a virtual assistant on screen and record how often people ask for a person. After reviewing conversations for errors, they adjust the content and expand the service.

Digital humans compared with robots and chatbots

A chatbot answers in text, a robot has a physical body and a digital human lives on a screen or in a headset with a human-like face. Each has strengths. Chatbots are cheap and fast, robots can act in the physical world and digital humans convey tone and body language without hardware. Many deployments combine them, for example a kiosk with a digital human that answers questions and a robot arm that hands over an item. The choice depends on whether the task needs emotion and trust, physical action or just information.

Measuring success

Organizations that deploy digital humans track task completion, customer satisfaction, handoff rate to human staff, response accuracy and cost per conversation. They also study how long people stay, whether they return and whether the avatar improves understanding compared with text alone. Because appearance can affect trust, teams run tests with diverse participants and compare several designs before launch. Accuracy audits of what the character says matter most, since a convincing face can make wrong answers more persuasive.

Future directions

Research is moving toward real-time digital humans that run on phones, that remember each user, respond to tone and facial expression and move with natural full-body gestures. Parallel work on watermarking, identity verification and disclosure rules aims to preserve trust. Many experts argue that the best digital humans will be those that are honest about being artificial, that serve clear purposes and that keep a person within reach when a situation calls for one.

FAQs

Is a digital human the same as an avatar? An avatar can be any digital representation. A digital human specifically aims for a realistic person that can converse.

Are digital humans real people? No. They are software characters, though some are modeled on real people with permission.

Will digital humans replace workers? They can handle routine tasks, but many roles need judgment and empathy, and companies often use them to assist human staff.

Key facts

  • Digital humans combine 3D graphics or neural rendering with speech, language models and animation to look and act like people.1
  • Early film examples such as the 2001 movie Final Fantasy: The Spirits Within and later work by Digital Domain pushed realistic CG faces.2
  • Realistic digital humans raise issues of consent, deception and emotional manipulation.3

Related concepts

Quick checkHow does a digital human differ from a simple chatbot?Show answer

A digital human adds a lifelike face, voice, expression and body language to conversation, so it communicates through a visible, human-like character and not only through text.

Sources

  1. Magnenat-Thalmann and Thalmann, Handbook of Virtual Humans, Wiley, https://onlinelibrary.wiley.com/doi/book/10.1002/0470023198
  2. Epic Games, MetaHuman Creator, https://www.unrealengine.com/en-US/metahuman
  3. Chesney and Citron, Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security, https://scholarship.law.bu.edu/faculty_scholarship/640/

Editorially reviewed by Specialty Digest Editorial TeamLast reviewed November 17, 2026Researched and drafted with AI assistanceReport an issue