How to Spot Fake Influencers: Check the Evidence

Start with a specific concern you can check. “The follower count jumped” is an observation. “The creator bought followers” is a conclusion that needs supporting evidence.

Quick Prompt: Review a creator authenticity concern

@Mysocial MCP

Edit your prompt Use Mysocial's held posts and my dated notes for [creator] to review [specific concern]. Separate observations, missing data and possible explanations. Cite each source and date, then give me one verification question and a provisional next step. A signal is not proof of fraud; do not invent a fake-follower percentage. Read existing evidence only. Draft here; do not save or contact anyone.

Review and send in your AI app.

Setup & access

Connect and select Mysocial before sending. Connect Mysocial →

Select your workspace and provide the known creator profile, concern and dated notes. Held external creator/post reads require Universe access and available coverage. They do not supply a complete follower audit or every post's comment text, private audience records or transcript. A comment count is not comment content. Missing data stays unknown. New collection can use an allowance; ask before collecting. Review the result yourself before making a partnership decision.

Result: A short evidence review: observation, source, possible explanation, unknown and next question.

Manual route: paste your dated source notes into your assistant, or use the review record below. Keep creator-supplied evidence separate from what you independently inspected.

Watch: how to detect fake influencers

Our original Mysocial video walks through the checks. Start here if you prefer a visual explanation, then use the screenshots below to examine a growth pattern.

What a useful authenticity check returns

Keep each finding short: observation → source and date → possible explanation → next question.

Five checks before a partnership

1. Put follower changes on a timeline

• Inspect: dated counts and the period covered. Place relevant posts, collaborations, promotions or giveaways on the same timeline where you have evidence.
• Question: what explains the change, and which source supports that explanation?
• Limit: a chart shows a recorded change, not its cause. Missing posts or history can leave the question unresolved.

See it: compare the growth curve with the daily counts

These are the original article’s Social Blade screenshots. Read them together: the chart helps you locate a change; the table gives you dates to investigate.

What to request: a link to the relevant post, collaboration or event, plus a dated platform record for the same period. These historical screenshots establish count changes; their cause remains unverified.

2. Compare engagement on the same basis

• Inspect: comparable platform, format and post age. Write the metric’s numerator and denominator.
• Question: is the unusual result still unusual against this creator’s comparable work?
• Limit: low engagement does not prove fraud; high engagement, saves or shares do not certify an audience. Do not apply a universal pass/fail percentage.

3. Read the comments you actually have

• Inspect: accessible comment text on a stated sample of posts. Keep short source-linked examples rather than guessing from the comment count.
• Question: are repeated phrases relevant to the content, and what context is missing?
• Limit: recurring commenters or emoji replies alone do not establish coordinated manipulation. A sparse or private profile is not automatically a bot.

4. Request evidence for the audience claim

• Inspect: a dated platform record relevant to the campaign’s audience requirement. Note the account, reporting period and whether it describes followers, viewers or reached accounts.
• Question: who supplied the record, what does it measure and can the important detail be checked?
• Limit: the creator’s location or language does not establish audience geography. Unavailable private analytics stay unknown; declining to share a record is not proof of fraud.

5. Check provenance before trusting a score

• Inspect: where the counts or estimate came from, when they were observed and what sample or method was used.
• Question: does the evidence support this particular claim about this account?
• Limit: an old screenshot, a live media kit and an automated estimate have different source limits. None should be presented as a complete independent authenticity audit without evidence for that claim.

Does Social Blade prove followers are fake?

Use a Social Blade chart, or any follower-history chart, to identify a change worth investigating. Record the dates and available context. The shape alone cannot establish that followers were purchased.

When a checker provides an estimated percentage, inspect what it classifies and which accounts it sampled. Do not present an unexplained estimate as the measured share of every follower who is a bot.

Worked example: investigate a follower jump

Original fictional teaching example. These invented notes are not an actual creator, customer result or verified fraud case.

A skincare brand is considering a creator. Its reviewer supplies:

• Note A: dated count snapshots show an increase during the review period.
• Note B: the creator says a collaboration caused the increase; the linked post and platform evidence have not been supplied.
• Note C: the reviewer has two recent posts and their comment counts, but no comment text or audience records.

A useful reviewed output

State: needs verification. Note A supports a recorded count change. Note B is a creator-supplied explanation to check. Note C cannot establish comment quality or audience authenticity.

Next question: “Can you share the collaboration post and a dated platform record for that period, including relevant discovery or audience information where available?”

Next action: inspect those sources before approving the partnership. If they remain unavailable, record the unresolved requirement and decide whether it prevents this campaign from proceeding.

Reject an AI answer that invents a bot percentage, calls the creator dishonest or says the collaboration is verified. None follows from the supplied notes.

Copy this review record

Creator / platform / review date:
Specific concern:
Observed evidence, source URL and date:
What the source can and cannot establish:
Possible explanation and who supplied it:
Missing record or access:
State: explained / needs verification / insufficient evidence
One next question:
Reviewer and next action:

Sources

Primary platform guidance

YouTube’s fake-engagement policy prohibits artificial inflation of views, likes, comments and other metrics. It also explains that traffic identified as artificial and subscribers identified as spam do not count toward the relevant totals. This describes platform enforcement, not a diagnosis of a particular creator.

Meta’s 2019 account of action against inauthentic Instagram activity describes enforcement against a service selling fake engagement. It is a dated example, not a current prevalence or detector-accuracy statistic.

Move from authenticity questions to campaign fit

Once the evidence is sufficient, use the separate influencer vetting checklist to assess the campaign brief, deliverables and remaining commercial questions.

Mysocial’s influencer discovery context can help you inspect held creator work with your assistant. You review the evidence and make the partnership decision.