In YouTube’s earliest years, paranormal videos were brief, blurry and supposedly accidental. Two decades later, ghost hunting has become a sophisticated entertainment industry built around feature-length investigations, specialist equipment, global fandoms and million-pound creator businesses. But has the evidence improved along with the production values?

When YouTube’s first 19-second video, Me at the Zoo, appeared in April 2005, it demonstrated the platform’s original appeal: ordinary people could record something, upload it and potentially show it to the world.

That same promise was irresistible to paranormal enthusiasts. A strange shape on a security camera, an unexplained voice on a recording or a shadow moving through an empty hallway no longer had to remain a local story. It could be uploaded, replayed, slowed down, brightened, debated and shared internationally.

YouTube did not invent the filmed ghost investigation. It inherited many of its conventions from television programmes such as Most Haunted, Ghost Hunters and Ghost Adventures: night-vision cameras, whispered commentary, electronic meters, dramatic histories and investigators challenging unseen entities to reveal themselves. What YouTube changed was who could participate.

A television network was no longer required. A haunted location, a camera and a willingness to enter a dark building could be enough.

Twenty-one years after YouTube’s launch, the platform’s paranormal genre is almost unrecognisable. Grainy clips have become feature-length documentaries. Amateur teams have become media companies. Comment sections have become mass evidence-review sessions. Ghost hunts have even moved into cinemas, merchandise businesses, subscription clubs and visitor attractions.

The evolution has been creatively impressive. Whether it has produced stronger evidence for ghosts is a more complicated question.

The first era: “Ghost caught on camera”

Early paranormal YouTube was dominated by the isolated evidence clip.

Titles usually promised something extraordinary: REAL GHOST CAUGHT ON CCTV, DEMON IN ABANDONED HOSPITAL or GHOST GIRL APPEARS BEHIND MAN. The videos were often short, context-free and heavily compressed. Many had passed through several websites or computers before reaching YouTube, making their original source difficult to establish.

This poor image quality was part of their power. Digital noise, reflections, compression artefacts and motion blur created ambiguous shapes that viewers could interpret in different ways. A few pixels might resemble a face. Dust illuminated close to a camera could become a glowing “orb”. A person in the distance could become transparent after being recorded through glass or processed repeatedly.

The audience was asked to perform a simple task: watch closely and decide whether the footage was real.

That made early ghost videos highly participatory. Viewers replayed individual frames, posted timestamps and argued over whether they had seen a spirit, an insect, a reflection or an actor. The comment section functioned as an informal public laboratory, although one without agreed procedures, access to original files or reliable information about how the footage had been produced.

The most compelling clips were often not those with the clearest evidence. They were those containing just enough ambiguity to sustain disagreement.

This remains one of the central rules of paranormal media: a completely clear image can be exposed as a person, prop or visual effect, while a vague image can remain mysterious indefinitely.

Monetisation changed the hunt

A major change arrived when uploading videos became a possible career.

YouTube introduced its Partner Program in 2007, allowing selected creators to receive a share of advertising revenue. The programme rewarded those who uploaded original material consistently and developed significant audiences.

That did not automatically make paranormal creators dishonest. Producing regular investigations involves genuine costs: transport, accommodation, insurance, location fees, cameras, batteries, sound equipment and editing time. Monetisation allowed committed teams to travel farther, purchase better equipment and produce more ambitious work.

It also created an unavoidable conflict.

A scientifically useful investigation may produce nothing. A commercially successful ghost-hunting channel must regularly produce videos that people consider worth clicking and watching.

The creator therefore enters a difficult position. A quiet night with no unusual events may be the most honest result, but “We Investigated a Building and Nothing Happened” is unlikely to perform as strongly as “The Demon Followed Us Home”.

This does not prove that evidence is fabricated. It means that the economic structure rewards dramatic interpretation. Ordinary noises become possible footsteps. Equipment malfunctions become “intelligent responses”. A frightened reaction becomes a thumbnail. Uncertainty may survive inside the episode, but certainty tends to dominate the title.

The pressure is not necessarily a deliberate command from YouTube. The platform says its recommendation system considers a mixture of signals, including watch time, satisfaction surveys, sharing, likes and dislikes, while demoting content classified as harmful or misleading in sensitive information categories.

Paranormal entertainment, however, occupies an awkward middle ground. A dramatic claim about a ghost is generally not treated in the same way as dangerous medical misinformation. Creators therefore compete largely through storytelling, personality, suspense and audience retention.

From the viral clip to the paranormal episode

In 2011, YouTube began integrating live-streaming capabilities directly into the platform. In 2012, it placed greater emphasis on watch time and longer viewing sessions rather than rewarding clicks alone.

These developments helped transform ghost content.

Instead of uploading one unexplained moment, investigators could construct a complete narrative:

1. Introduce the location and its history.
2. Interview witnesses.
3. Explore the building in daylight.
4. Establish alleged paranormal “hotspots”.
5. Begin the night investigation.
6. Escalate from quiet observation to equipment sessions.
7. End with the strongest apparent evidence.

This was much closer to television, but YouTube allowed episodes to be longer, looser and more personal. Creators could spend 40 minutes, an hour or sometimes several hours inside one location. They could also upload deleted scenes, evidence reviews and behind-the-scenes footage.

The investigator’s personality became as important as the haunting. Audiences were no longer simply watching a building. They were spending time with a familiar group of people.

That relationship became the genre’s greatest asset. A viewer might doubt a particular recording while still trusting the creators’ sincerity. Fear, friendship, humour and vulnerability created emotional credibility even when the physical evidence remained inconclusive.

The ghost-hunting technology arms race

As paranormal channels grew, their equipment bags expanded.

Early investigations often relied on digital cameras, audio recorders and basic electromagnetic-field meters. Modern episodes may include thermal cameras, motion sensors, REM pods, laser grids, vibration detectors, radio-sweep “spirit boxes”, environmental data loggers and devices that convert sensor readings into words.

The equipment creates an appearance of scientific procedure. Screens display numbers. Lights respond. Graphs change. Devices produce voices. Investigators conduct what they call experiments.

The difficulty is that an instrument can be scientific without its paranormal interpretation being scientifically justified.

EMF meters

An electromagnetic-field meter measures electromagnetic fields. These may be produced by wiring, electrical appliances, radio transmitters, mobile phones and other ordinary sources.

Ghost hunters frequently interpret an unexplained rise as evidence that a spirit is present or drawing energy from the environment. However, no accepted mechanism has established that a deceased human consciousness should create an electromagnetic field, or that a particular reading distinguishes a ghost from conventional interference.

A spike is evidence that the meter detected something within its operating range. Determining the source requires controlled testing, not merely asking the spike whether it is a ghost.

Electronic voice phenomena

Electronic voice phenomena, or EVP, are apparent voices discovered in recordings made during investigations. Some are clear enough to attract attention without prompting. Many are extremely faint and are presented with subtitles telling the audience what to hear.

This is important because human perception is strongly influenced by expectation.

Research into auditory pareidolia shows how people can perceive meaningful words in ambiguous or random sound. In one study discussed in PLOS One, participants exposed to paranormal suggestions were more likely to report voices in unclear audio. Previous experiments cited by the researchers found that 36 per cent of participants reported hearing voices in random noise after first watching information about EVP. The same paper found remarkably little agreement between listeners and the words paranormal investigators had assigned to EVP clips.

This does not establish that every EVP has an ordinary explanation. It does demonstrate why subtitles, repeated listening, noise reduction and leading questions can unintentionally create apparent messages.

A stronger test would allow multiple independent listeners to describe the recording without being shown subtitles, questions or location history. Their answers could then be compared. If one listener hears “leave this house”, another hears “Steven’s out” and a third hears only static, the evidential value is weak.

Spirit boxes and the Estes Method

A spirit box rapidly scans radio frequencies, producing static and fragments of broadcasts. Investigators believe spirits may select or manipulate these fragments to answer questions.

The Estes Method attempts to reduce some forms of cueing. One participant listens through headphones, sometimes while blindfolded, and repeats words heard from the scanning radio. Another participant asks questions that the listener supposedly cannot hear.

It is a more interesting procedure than an ordinary group spirit-box session because it separates the questioner from the listener. It does not, however, remove the central problem: the device deliberately generates fragments of human speech.

Given enough questions, words and time, occasional matches are likely. Investigators may also remember impressive responses while forgetting the many irrelevant ones. This is a form of selection bias. A genuine test would record every question, every spoken fragment, the timing between them and predefined rules for deciding what counts as a match.

SLS and Kinect figures

Structured-light or SLS cameras are used to map human-shaped figures into an environment. Paranormal investigators sometimes interpret a stick figure appearing in an empty room as an invisible entity.

Many of these systems originated from Microsoft’s Kinect technology, which was designed to identify human bodies for gaming. The software actively attempts to construct a skeleton from depth information. Furniture, railings, shadows and irregular surfaces can therefore be misidentified as body parts.

In other words, the device may display a human-like figure because it is designed to search for one, not because it has detected a spirit. Experts examining the technology have described its false skeletal figures as a predictable consequence of the system’s body-recognition process.

The more visually impressive the stick figure appears, the easier it is to forget that viewers are watching a software interpretation rather than a photograph of an entity.

The BuzzFeed Unsolved turning point

One of the most significant changes in paranormal YouTube arrived with BuzzFeed Unsolved: Supernatural.

Launched in 2016, the series paired Ryan Bergara, an enthusiastic believer, with Shane Madej, an openly sceptical co-host. Rather than hiding disagreement, the programme made disagreement its central format.

Ryan interpreted locations through folklore, witness testimony and personal experience. Shane challenged assumptions, mocked demons and regularly remained unimpressed by supposed evidence. Their chemistry allowed viewers to occupy either position without being excluded.

By 2019, BuzzFeed reported that the wider Unsolved franchise had accumulated hundreds of millions of views and billions of minutes of watch time. The programme ended its original run in 2021 after almost six years, with each final investigation accompanied by a question-and-answer “postmortem”.

Its influence was deeper than its viewing figures.

Unsolved demonstrated that scepticism did not have to destroy paranormal entertainment. It could improve it. Shane’s refusal to accept every noise as evidence gave the series tension and credibility. Ryan’s willingness to enter frightening situations prevented the scepticism from becoming distant or sneering.

The viewer was not told what to believe. The viewer was invited into an ongoing argument.

That believer–sceptic structure remains one of the healthiest models available to paranormal media. It respects extraordinary experiences without automatically treating every interpretation as factual.

Ryan and Shane later continued their paranormal work independently through Watcher’s Ghost Files, illustrating another major development: creators were no longer dependent on large digital publishers. Successful personalities could establish their own production companies, retain their audiences and build television-scale programmes directly on YouTube.

The influencer investigation

While BuzzFeed Unsolved refined the believer–sceptic documentary, other creators fused ghost hunting with exploration, challenges and influencer culture.

Sam Golbach and Colby Brock are among the clearest examples. Their work evolved from stunts and abandoned-building exploration into large-scale paranormal investigations. Academic analysis has placed channels like theirs within a form of “virtual dark tourism”, allowing viewers to experience prisons, hospitals, historic houses and sites associated with death or tragedy from home.

Their investigations increasingly resemble event films. Episodes use cinematic drone footage, elaborate sound design, multiple cameras, invited mediums, celebrity guests and recurring Halloween seasons. Locations become franchises, with investigations spread across several videos or an entire week.

The business surrounding the videos has also expanded. Sam and Colby have developed clothing, memberships, live experiences and theatrical releases. Reporting on their company showed how paranormal content can support a diversified creator business stretching far beyond advertising revenue.

This represents an extraordinary change from the anonymous ghost clip.

The modern paranormal creator may be presenter, investigator, filmmaker, brand owner, tourism operator and executive producer at the same time.

It also intensifies the problem of evidence.

When an investigation employs guests, dramatic music, fast editing and a major production crew, the audience cannot easily determine what happened continuously, what occurred off camera or how much material was discarded. Even completely sincere creators are producing an edited story rather than a controlled scientific record.

The question is no longer simply, “Could this clip be fake?”

It is also, “Can a commercial entertainment production ever provide enough information to test its own conclusion?”

The audience becomes part of the investigation

YouTube’s greatest innovation may not be its cameras or equipment. It is the audience.

Viewers freeze frames, isolate sounds and identify background movements missed by the production team. They submit personal stories, recommend locations, fund investigations and sometimes travel to places featured in videos.

Some series invite audience members to submit their own alleged evidence before an investigation. Livestream viewers can report noises or shapes in real time. After release, sceptical channels produce detailed breakdowns, recreating effects and testing equipment.

This participation can improve accountability. A misleading edit may be noticed. A hidden crew member may be spotted in a reflection. A supposedly mysterious sound may be identified as plumbing, wildlife or radio interference.

Participation can also amplify weak evidence. Once one viewer circles a dark patch and labels it a face, thousands of others may begin to see the same face. Repetition transforms a suggestion into a community belief.

Research published in 2024 examined 50 highly viewed YouTube videos about hauntings and UFOs. Most presented paranormal claims, while scientific sources appeared in almost none of the haunting videos. A survey conducted as part of the study found an association between YouTube use and belief in hauntings, even after accounting for several other variables. The research does not prove that YouTube directly causes belief, but it suggests the platform can help reinforce paranormal interpretations.

The genre therefore does more than entertain existing believers. It helps establish what paranormal investigation looks like, what counts as evidence and which explanations viewers consider plausible.

The pandemic, gaming and the paranormal revival

Interest in ghost content grew again during and after the COVID-19 pandemic.

Periods of isolation, uncertainty, bereavement and heightened awareness of mortality created fertile ground for supernatural storytelling. People spent more time at home, consumed more online entertainment and sometimes became newly attentive to unfamiliar noises in their surroundings.

The multiplayer game Phasmophobia, released in 2020, introduced a younger audience to a gamified version of paranormal investigation. Players use recognisable ghost-hunting equipment, collect evidence and identify different entities. Streamers and YouTubers helped turn the game into a cultural phenomenon, while some players moved from virtual investigations towards real-world paranormal groups.

British ghost-hunting communities have reported a broader rise in “have-a-go” investigators, commercial ghost nights and paranormal tourism, supported by YouTube and TikTok exposure.

The exchange now works in both directions. Television inspired early YouTube ghost hunters. YouTube inspired games and independent creators. Games introduced new investigators to YouTube. Successful YouTubers then moved into television-scale productions and cinemas.

Ghost hunting has become a connected entertainment ecosystem.

The Shorts era: fear in under a minute

The arrival of YouTube Shorts added another layer.

YouTube began testing Shorts creation tools in 2020. The format was designed for rapid, mobile viewing and easy discovery.

For paranormal creators, Shorts function as trailers, highlights and evidence clips. A two-hour investigation can be condensed into 30 seconds containing a scream, a moving object and a device apparently saying somebody’s name.

This is effective promotion, but it removes context.

Viewers may not see the ten minutes of silence before the event, the investigators walking near the equipment or the later attempt to recreate the sound. The most dramatic interpretation reaches millions of viewers without the qualifications contained in the full episode.

Short-form paranormal media risks returning the genre to its earliest form: an isolated, unexplained clip detached from the conditions under which it was recorded.

Only now the image is sharper, the editing faster and the distribution algorithm far more powerful.

Has the evidence actually improved?

Technically, yes.

Modern creators have access to better low-light cameras, clearer audio, multiple viewpoints, precise timestamps and enormous storage capacity. Raw footage can be preserved. Environmental conditions can be measured. Investigations can be livestreamed. Independent analysts can examine material immediately.

In principle, this should make paranormal claims easier to test.

Yet improved technology does not automatically produce improved methodology.

A 2020 review of environmental research into allegedly haunted locations found a relatively small and methodologically inconsistent body of literature. Studies examining factors such as lighting, temperature, air quality, infrasound and electromagnetic fields produced outcomes that were mostly null, weak or inconsistent. The authors called for more precise measurement, proper control conditions and cooperation between researchers with differing beliefs.

Those recommendations highlight the gap between an investigation and an experiment.

An investigation asks, “What happened here tonight?”

An experiment asks, “Under clearly defined conditions, does this effect occur more frequently than chance, error or ordinary environmental activity would predict?”

Most YouTube ghost hunts are excellent at the first question. Very few are designed to answer the second.

What a genuinely rigorous YouTube ghost investigation would involve

A more credible investigation would not need to abandon atmosphere, storytelling or curiosity. It would need stronger controls.

Investigators could begin by recording environmental baselines before discussing the building’s ghost stories. Electrical sources, radio signals, airflow, loose windows, plumbing, wildlife and building movement could be documented in daylight.

Equipment should be tested openly. Viewers should be shown what causes a K2 meter, motion sensor or REM pod to activate under normal conditions. Devices should be kept away from phones, radios and camera transmitters unless interference is part of the test.

Audio should be reviewed by independent listeners who do not know the questions being asked. EVP subtitles should appear only after viewers have first heard the untreated recording. Original audio files should be provided when privacy and location agreements allow.

Every response from a spirit-box session should be logged, including irrelevant words and failed answers. Criteria for a successful response should be established before the session rather than invented afterwards.

Multiple fixed cameras should cover investigators and equipment simultaneously. Continuous timestamps would make it harder for editing to create a misleading sequence.

Negative results should be included. If a device remains silent for three hours, that information is essential for assessing the significance of its one activation.

Most importantly, an apparently successful result should be repeated. A response that occurs once may be interesting. A response that occurs reliably under controlled conditions would be far more important.

These practices would not prove that ghosts do not exist. They would help distinguish genuinely unexplained events from the enormous number of ways people, buildings and electronic devices can produce strange results.

The coming problem: AI ghosts

The next stage of paranormal YouTube will face a challenge greater than fishing line, hidden speakers or selective editing.

Realistic synthetic video and audio can now create people, voices and events that never existed. A convincing apparition no longer requires advanced visual-effects training. An alleged EVP can be generated with a cloned voice. Security-camera footage can be altered while retaining the imperfect appearance audiences associate with authenticity.

YouTube introduced disclosure requirements for realistic altered or synthetic content in 2024, including footage that could mislead viewers about real people, places or events.

Policies alone will not solve the problem. Paranormal creators who care about credibility may need to preserve original files, publish metadata, document chains of custody and use cameras capable of generating verifiable authenticity information.

Ironically, the future of ghost hunting may depend less on obtaining more mysterious footage and more on proving that the footage has not been manipulated.

Scepticism is not the enemy of wonder

Paranormal YouTube is sometimes judged by two equally unhelpful extremes.

One side assumes that every unexplained event is supernatural. The other assumes that every investigator is fraudulent and every witness foolish.

Neither approach reflects reality.

People can sincerely misinterpret ordinary events. Equipment can behave unpredictably. Memories can change. Fear can influence perception. Editors can unintentionally exaggerate a sequence. Investigators can become emotionally invested in a theory without consciously deceiving anyone.

There are also deliberate hoaxes. The history of paranormal media contains actors, props, manipulated photographs and staged events. But exposing individual hoaxes does not explain every unusual experience, just as failing to explain an experience does not establish that a ghost caused it.

Healthy scepticism means withholding a conclusion until the evidence supports it. It asks whether a normal explanation has been properly tested. It distinguishes “unexplained” from “supernatural”. It remains willing to change position when better information appears.

Most importantly, scepticism need not strip ghost hunting of its value.

Paranormal investigations preserve local folklore, introduce audiences to neglected historic buildings and create communities around mystery, mortality and storytelling. They can be funny, frightening and emotionally meaningful even when they fail to demonstrate an afterlife.

The strongest creators understand that uncertainty is not a weakness. It is the reason viewers continue watching.

From “broadcast yourself” to building haunted worlds

YouTube began as a place for brief, personal uploads. Its paranormal genre followed the same journey as the platform itself.

It moved from isolated clips to regular channels, from amateur recordings to professional productions, from advertising revenue to complete creator businesses. Ghost hunters became hosts, filmmakers and entrepreneurs. Viewers became fans, investigators, critics and customers.

The production quality has improved beyond anything early uploaders could have imagined. The central question, however, remains unchanged:

What exactly are we looking at?

After two decades, YouTube has not settled the existence of ghosts. What it has done is transform the search for them into one of the internet’s most adaptable forms of storytelling.

The future of the genre will depend on whether creators continue escalating the spectacle or recognise that the greatest unexplained event would not need dramatic music, a shocked thumbnail or a glowing gadget.

It would need evidence strong enough to survive believers, sceptics and the pause button.