AI influencers can go viral when they present something an ordinary human cannot, such as a singing cat, a character dancing in an unnatural way, or a situation that deliberately breaks logic. However, attention created by absurdity often moves on quickly. A character can last only when it grows from a visual surprise into intellectual property, or IP, with an identity, a story, and a reason for people to keep following it.
Viral characters such as Jean Phil, as mentioned in the original material for this article, demonstrate that appeal. Viewers are not only looking for perfect visuals. They are looking for something strange, funny, surprising, and easy to share with someone else.
The most important question for a brand is not, “Can AI make viral content?” A more useful question is, “After the surprise fades, will people still remember the character, wait for its story, and take an action that matters to the business?”
Why do people enjoy absurd AI content?
Absurd content works because it violates expectations in a way people can understand within seconds. An audience needs no long explanation to recognize that a singing cat, talking food, or a digital human behaving strangely is unusual.
Five main forces drive its appeal:
- Surprise. An unexpected visual makes people stop scrolling.
- Humor without lengthy context. Strangeness can become the joke immediately.
- Shareability. People send the content with a message such as, “You have to see this.”
- Participation. Audiences can create their own versions, remixes, comments, or theories.
- Fast production. AI allows many variations to be produced before the trend loses momentum.
Hootsuite's 2026 social media trends report describes the dominance of “chaos culture” and absurd humor among Generation Alpha. The report also offers an important warning: AI has become a common production tool, but authenticity remains a differentiator. Strangeness can attract attention, but a lasting relationship still requires a human touch, a point of view, and a story that feels intentional.
Will the absurd AI influencer trend last?
The format will last, but most viral characters will not. AI's ability to create impossible realities will not disappear. In contrast, a particular character, joke, voice, or format can lose its appeal once audiences see too many versions of the same idea.
Tung Tung Tung Sahur and the wave of other AI slop characters demonstrate a recognizable pattern. A character appears, gets copied at scale, is transformed into many variations, and gradually loses its shock value as timelines fill with imitations. Saying that it has completely “disappeared” would be too absolute. Memes, communities, and derivatives can remain active. However, peak public attention does not automatically become a long term business asset.
The same can happen to an AI influencer. A consistent face and distinctive visuals may be enough to earn the first click. They are not enough to make people return in the following weeks or months.
A trend's lifespan cannot be reduced to one universal number. A microtrend can surge and decline within days or weeks. A format that keeps evolving can last for months. An IP that successfully builds a world, conflict, and community can live for years. The calendar is therefore not the best measure. Signs of audience fatigue are far more useful.
Signs that a trend is starting to weaken include:
- Share and save rates fall even as posting volume increases.
- Comments discuss the technology more often than the character.
- New versions no longer create surprise.
- Follower growth stops after one or two major posts.
- Audiences recognize the format but cannot remember the character's name.
- Viewers do not become community members, subscribers, leads, or buyers.
Going viral is not the same as becoming an IP
Viral content is an event, while IP is a system that can keep generating stories and value. One video can receive millions of views without creating a character that can endure.
An absurd character starts to show IP potential when it has:
- A memorable and searchable name.
- A consistent silhouette, voice, color palette, or habit.
- Desires, weaknesses, and conflicts that can be developed.
- A story world that supports many episodes.
- Documented visual and behavioral rules.
- Room for the audience to participate.
- Clear rights to use the assets, music, models, and training material.
- Opportunities to expand into products, collaborations, licensing, or other experiences.
The distinction is simple. “A dancing AI cat” is a format anyone can imitate. “A specifically named cat that always dances when it lies, lives in a defined world, and has relationships with other characters” is already moving toward a story property.
A brand should therefore protect more than a prompt. Prompts can be copied and models can change. What needs protection is the name, character design, character bible, behavioral patterns, story world, source assets, relevant copyrights, and relationship with the community.
What can we learn from Roku Studio's Instagram experiment?
Roku Studio's AI influencer experiment on Instagram can serve as an initial test, but it is not enough to determine viral potential. Based on the public preview checked on October 7, 2026, the post dated October 6, 2026 showed 8 likes and 0 comments.
Those figures cannot yet be used to calculate cost per result or judge success because the following important data is not publicly visible:
- Plays or impressions.
- Unique account reach.
- Average watch time and completion rate.
- Shares and saves.
- Profile visits and follower growth.
- Organic or paid distribution sources.
- The number and value of business actions after viewing.
Eight likes also do not mean that the experiment failed. The measurement took place roughly one day after publication, and there is no denominator in the form of reach. At the same time, the available data provides no basis for describing the content as viral.
The next step should not be to simply increase the budget behind one post. The experiment should become a series with several controlled variations:
- Use the same character for at least 30 days.
- Test three types of hook: visual surprise, character conflict, and local humor.
- Create short episodes instead of collections of random visuals.
- Preserve the character's identity while changing situations and conflicts.
- Measure share rate, save rate, completion rate, and follower growth per 1,000 people reached.
- Promote only the content that already shows the strongest organic signals.
How much does it cost to make an AI influencer go viral?
No price can guarantee virality. A budget can buy experiment quality, production volume, distribution, and learning speed. Organic virality remains an outcome of audience response and platform algorithms.
The following estimates are Roku Studio planning models, not official platform rates. They must be adjusted for character complexity, video quality, tool licensing, revision volume, human talent, and target market.
Scenario 1: A 30 day experiment
Realistic budget: IDR 25 million to IDR 45 million.
A reasonable scope includes:
- Lightweight character development and a visual guide.
- 8 to 12 short videos.
- 8 to 12 supporting pieces such as carousels, still images, or stories.
- Copywriting and basic community management.
- IDR 8 million to IDR 15 million in media testing.
- A performance report and a decision about which concept to continue.
This scenario is suitable for answering one question: does the character show a strong enough early signal to justify further testing?
Scenario 2: IP validation over 90 days
Recommended budget: IDR 100 million to IDR 150 million.
The suggested scope includes:
- A character bible, personality, world, and story arcs.
- 15 to 20 master references to maintain consistency.
- 36 to 45 pieces of content over 90 days.
- 12 to 18 short videos as the main assets.
- Community management and social listening.
- Paid media tests on the strongest content.
- Limited seeding through creators or community accounts.
- Basic checks on the name, trademarks, assets, and usage rights.
- An evaluation dashboard for attention, retention, community, and business actions.
This scenario does not buy virality. The budget buys enough trials for the team to distinguish a strong idea from the luck of one post.
Scenario 3: An aggressive 90 day launch
Realistic budget: IDR 250 million to IDR 500 million or more.
The scope can include more intensive video production, broader creator seeding, collaborations, paid media, adaptation across platforms, daily community management, and IP asset development. This scenario becomes rational only when early testing shows that audiences remember and follow the character.
Cost simulation for reaching one million people
One million paid impressions are not one million instances of quality attention, and they do not prove virality. However, a paid media simulation can help establish budget limits.
If a team uses a planning CPM assumption of IDR 20,000 to IDR 60,000, one million impressions would require about IDR 20 million to IDR 60 million in media spending. If the goal is close to one million unique people at an average frequency of 1.5, the campaign would need about 1.5 million impressions, or IDR 30 million to IDR 90 million.
The formula is:
Media cost = target impressions ÷ 1,000 × CPM
With production at IDR 30 million to IDR 70 million, creator seeding at IDR 15 million to IDR 100 million, and media at IDR 30 million to IDR 90 million, the total campaign cost for pursuing large scale reach can fall between IDR 75 million and IDR 260 million. That figure still does not guarantee organic conversation, shares, media coverage, or community growth.
Actual CPM can be lower or higher because platforms use auction systems. Target audience, country, season, objective, creative quality, and advertiser competition all affect the result. The CPM figures above should therefore be treated as planning assumptions to test through a small campaign, not as promised prices.
Metrics that are more honest than likes
An AI influencer's success should be judged by retention and business value, not views alone. Likes are easy to see, but they often fail to show whether the content was remembered or led to action.
Use four layers of metrics:
1. Attention
- Reach and impressions.
- Thumb stop rate, or the ability to stop someone from scrolling.
- Average watch time.
- Completion rate.
2. Resonance
- Shares per 1,000 people reached.
- Saves per 1,000 people reached.
- Comments that mention the character's name.
- Memes, remixes, or fan content created by the audience.
3. Retention
- Net follower growth.
- Percentage of returning viewers.
- Performance of later episodes compared with the first episode.
- Direct traffic and searches for the character's name.
4. Business impact
- Profile visits and website clicks.
- New leads or subscribers.
- Cost per lead and cost per customer.
- Revenue, contribution margin, and repeat purchases.
- Collaboration or licensing inquiries.
Roku Studio's article on viral content and sales explains the same problem from a business perspective. Broad reach does not automatically create sales. To understand the increasingly crowded competition for attention, also read AI content competition in the attention economy.
How to keep an absurd character from dying quickly
Strangeness should be the entrance, not the character's entire substance. After an audience stops for an unusual visual, it needs a reason to follow the next episode.
A more durable strategy is to:
- Build a repeatable premise. Each episode can be different while still feeling as if it comes from the same world.
- Give the character a goal and a problem. Audiences follow conflict, not technology.
- Use absurdity consistently. Strangeness needs rules so it does not feel like random output.
- Include human observations. Humor about work, relationships, status, money, or local habits makes fantasy feel relevant.
- Invite participation. Let the audience choose a decision, propose a name, or suggest the next situation.
- Test quickly, but curate strictly. AI accelerates production. It does not remove the need for art direction.
- Stop concepts that survive only on novelty. If people only say, “AI is so advanced now,” the character has not become an IP.
Production quality also needs careful control. As explained in Roku Studio's article about AI slop, prompting, and art direction, strong generative work still requires a concept, visual direction, iteration, and human quality control.
Conclusion
AI influencers and absurd content will remain part of social media because people enjoy surprise, humor, and impossible worlds. However, not every viral character will last. Most will fade once novelty runs out and imitations flood the timeline.
For brands, the most rational choice is not to pursue one viral video with the largest possible budget. Start with a measured experiment, then continue only if audiences remember the character's name, wait for the next episode, participate organically, and take actions that create value.
For a serious project, the recommendation is a 90 day pilot with a budget of IDR 100 million to IDR 150 million. The goal is not to guarantee virality. The goal is to prove whether a character can evolve from AI slop that passes through the timeline into an entertainment IP worth building.
To design the character, production system, and distribution test, discuss your project with Roku Studio.
Frequently asked questions
Can an AI influencer still go viral?
Yes. An AI influencer can still go viral with a strong visual hook, story, humor, or cultural relevance. However, using AI alone is no longer a differentiator because the same tools are available to many creators.
How much does it cost to create an AI influencer?
A 30 day experiment can be planned at IDR 25 million to IDR 45 million. A 90 day IP validation is more realistic at IDR 100 million to IDR 150 million. An aggressive launch may require IDR 250 million to IDR 500 million or more.
Can a large budget guarantee viral content?
No. A budget can buy production, media, creator seeding, and experiments, but it cannot guarantee an organic response. One million paid impressions are also not the same as one million people remembering or liking the character.
What is the difference between a viral character and intellectual property?
A viral character receives short term attention. IP has a consistent identity, expandable stories, a community, assets that can be protected, and economic potential through collaborations, products, or licensing.
Which metrics should be checked beyond likes?
Check watch time, completion rate, share rate, save rate, comments that mention the character's name, follower growth, returning viewers, searches for the character, clicks, leads, and sales.
Sources and methodology notes
- Roku Studio, AI influencer experiment on Instagram. The public preview was checked on October 7, 2026. The visible data included the post date, likes, and comments. Reach, plays, shares, saves, and internal account data were not available.
- Hootsuite, Social Media Trends 2026. Used for context on chaos culture, AI use, and the importance of authenticity.
- Shopify, Influencer Pricing 2026. Used as a reference showing that creator costs vary widely by platform, audience quality, engagement, format, usage rights, and exclusivity.
- WeRSM, virtual influencer engagement report. This report cites 2019 HypeAuditor data. Its figures are historical and are not used as evidence of current performance or profitability.
- All Indonesian rupiah estimates in this article are planning models, not official rates or guarantees of results.