Two fabricated Instagram accounts, built entirely on purchased followers, landed four paid brand deals in a Mediakix test. Nobody caught them. That is the gap AI authenticity scoring is meant to close: a repeatable, auditable way to verify creator engagement before budget moves. Here is what these scores measure, where they mislead, and how to pilot one.
Why follower counts stopped being a safe proxy
In a Mediakix test reported by Digiday, two fake Instagram accounts with purchased followers and fake engagement secured four paid brand deals. It cost between $3 and $8 to buy about 1,000 followers, and the platform did not flag the rapid growth. Fakery was cheap, and detection was left to the buyer.
Today the fakes are better. Bot networks mimic human comment patterns. Pods trade likes on schedule. Engagement farms rent real devices. A creator can look healthy on a media kit and still deliver an audience that never buys.
So the question changes. Not “how many followers?” but “how much of this attention is real, and can I prove it?”
What an authenticity score actually measures
An authenticity score is a composite. Vendors weight the inputs differently, but most models draw on the same families of signals:
- Audience quality: share of followers that look like active, geographically plausible humans versus dormant, mass-following or newly created accounts.
- Growth shape: follower curves that jump in steps, then flatten, usually signal purchased batches.
- Comment semantics: repeated generic phrases, emoji-only replies, and comments unrelated to the post content.
- Engagement ratios: likes, comments, saves and shares relative to reach, benchmarked against creators of similar size and niche.
- Network behavior: clusters of accounts that interact with each other far more than with anyone else, a pod fingerprint.
The output is usually a number from 0 to 100, sometimes with a risk tier. Treat that number as a triage tool. It tells you where to look, not what to conclude.
A score is a hypothesis about an audience, not a verdict. The brands that get value from it use it to decide who gets a manual review, not who gets rejected.
The standards question: who sets the bar?
Here is the uncomfortable part. There is no universal authenticity standard. Two vendors can score the same creator 62 and 88, and both can be defensible, because each defines “authentic” differently.
What does exist is a regulatory floor. The FTC’s final rule on reviews and testimonials prohibits anyone from selling or buying fake indicators of social media influence, such as followers or views generated by a bot or hijacked account. The rule’s text sits in section 465.8 of the Federal Register notice, and it took effect in October 2024.
The same release notes the rule lets the agency seek civil penalties against knowing violators. Read that word, “knowing”, carefully. If your vendor flags a creator as high-risk and you buy anyway, your file shows you knew.
That cuts both ways. A documented scoring process becomes evidence of diligence. An ignored score becomes evidence of the opposite.
Disclosure is the adjacent standard
Authenticity is not only about bots. In the UK, the ASA, CAP and CMA have also published a new edition of their joint influencer guide, which now covers fake reviews. A review must reflect the influencer’s genuine experience, or it breaches both the law and the CAP Code.
Why does this matter for scoring? Because an engagement audit that ignores content authenticity is half an audit. A creator with a real audience who reads a script about a product they never used still creates risk. Your scoring workflow should have a human step for that. Briefs built around evidence rather than scripts help here, and our piece on creator briefs covers how.
Where scoring models go wrong
I have seen teams treat a vendor dashboard as gospel. Four failure modes come up repeatedly.
- False positives on niche creators. Small, tight-knit communities can look like pods because members genuinely interact with each other.
- Regional blind spots. Models trained mostly on English-language audiences misread comment patterns in other markets.
- Platform data access. Scores depend on what APIs expose. When access narrows, the model guesses.
- Black-box weighting. If a vendor will not explain why a creator dropped 15 points, you cannot defend the decision to a client or a regulator.
None of these is a reason to skip scoring. They are reasons to interrogate it.
How to vet a scoring vendor
Ask for specifics, and expect uncomfortable answers.
- Which signals feed the score, and can you see them per creator?
- How often is the score refreshed? A snapshot from three months ago is stale.
- What is the documented false-positive rate, and how was it tested?
- Can you export raw findings for your own records?
- How does the tool handle creators across TikTok, YouTube and Instagram, where engagement norms differ?
Our guide to vendor vetting for fraud detection goes deeper on contract terms and audit rights. Read it before you sign anything.
Run a pilot, with a control
Do not roll a score across your roster on day one. Pick one campaign or one tier of creators, and test whether the score predicts business outcomes.
A simple design works. Score all candidates. Include some high scorers, some middling, and a few flagged ones you would otherwise accept. Track downstream results: click quality, conversion, promo-code redemptions, and cost per acquisition. Then check whether score and outcome move together.
The logic is the same as any incrementality test. If you want the methodology, see how holdouts keep AI-driven decisions honest.
The best authenticity score is the one that predicts your own conversions. A score that cannot beat your current vetting process is just another line item.
Connect scores to attribution
A score sitting in a spreadsheet changes nothing. It earns its keep when it sits next to performance data, so you can see that a creator with a mediocre score drives outsized sales or that a pristine one does not. That requires joined-up data. Many teams still run siloed creator attribution, which makes this comparison painful. Fix the identity layer first, or your pilot results will be noise.
A practical workflow for brand teams
- Screen: run every shortlisted creator through the automated score.
- Triage: set thresholds for auto-advance, manual review and decline. Document them.
- Review: a human checks audience demographics, comment samples and growth history for anything in the middle band.
- Contract: write authenticity warranties and audit rights into agreements, including remedies if purchased engagement is found.
- Monitor: rescore during the campaign. Fraud often appears mid-flight, after a creator sees a deal land.
Keep a log of each decision. If a regulator or client asks why you hired someone, you want a dated record, not a recollection.
Next step
This week, pull your last five creator contracts and check whether any clause covers purchased engagement. Then select one upcoming campaign as a scoring pilot, with outcome metrics agreed before the first score is run.
FAQs
What is AI authenticity scoring?
It is a model-generated rating of how genuine a creator’s audience and engagement appear. It combines signals such as follower quality, growth patterns, comment content and engagement ratios into a single score or risk tier that brands use to prioritize reviews.
Is there an official standard for authenticity scores?
No. Vendors define and weight signals differently, so scores are not comparable across tools. The closest thing to a standard is regulatory: the FTC rule bans buying or selling fake social media indicators, which sets a legal floor for what you may purchase or knowingly benefit from.
Can a high authenticity score guarantee campaign results?
No. A real audience can still be a poor fit for your product. Authenticity tells you the attention is human. Relevance, creative quality and offer strength determine whether it converts.
How often should creators be rescored?
Score at selection, then again during and after the campaign. Audience quality can shift quickly, and mid-campaign purchases of engagement are a known risk.
Should brands reject creators below a certain score?
Use tiers instead of a single cutoff. Auto-advance clear passes, send the middle band to manual review, and decline only when a human confirms the red flags. This limits false positives, especially for niche creators.
What should be in a creator contract about fake engagement?
Include a warranty that no followers, views or engagement were purchased or generated by bots, a right to audit analytics, and defined remedies such as fee reduction, repayment or termination if violations are found.
Top Influencer Marketing Agencies
The leading agencies shaping influencer marketing in 2026
Agencies ranked by campaign performance, client diversity, platform expertise, proven ROI, industry recognition, and client satisfaction. Assessed through verified case studies, reviews, and industry consultations.
Moburst
-
2
The Shelf
Boutique Beauty & Lifestyle Influencer AgencyA data-driven boutique agency specializing exclusively in beauty, wellness, and lifestyle influencer campaigns on Instagram and TikTok. Best for brands already focused on the beauty/personal care space that need curated, aesthetic-driven content.Clients: Pepsi, The Honest Company, Hims, Elf Cosmetics, Pure LeafVisit The Shelf → -
3
Audiencly
Niche Gaming & Esports Influencer AgencyA specialized agency focused exclusively on gaming and esports creators on YouTube, Twitch, and TikTok. Ideal if your campaign is 100% gaming-focused — from game launches to hardware and esports events.Clients: Epic Games, NordVPN, Ubisoft, Wargaming, Tencent GamesVisit Audiencly → -
4
Viral Nation
Global Influencer Marketing & Talent AgencyA dual talent management and marketing agency with proprietary brand safety tools and a global creator network spanning nano-influencers to celebrities across all major platforms.Clients: Meta, Activision Blizzard, Energizer, Aston Martin, WalmartVisit Viral Nation → -
5
The Influencer Marketing Factory
TikTok, Instagram & YouTube CampaignsA full-service agency with strong TikTok expertise, offering end-to-end campaign management from influencer discovery through performance reporting with a focus on platform-native content.Clients: Google, Snapchat, Universal Music, Bumble, YelpVisit TIMF → -
6
NeoReach
Enterprise Analytics & Influencer CampaignsAn enterprise-focused agency combining managed campaigns with a powerful self-service data platform for influencer search, audience analytics, and attribution modeling.Clients: Amazon, Airbnb, Netflix, Honda, The New York TimesVisit NeoReach → -
7
Ubiquitous
Creator-First Marketing PlatformA tech-driven platform combining self-service tools with managed campaign options, emphasizing speed and scalability for brands managing multiple influencer relationships.Clients: Lyft, Disney, Target, American Eagle, NetflixVisit Ubiquitous → -
8
Obviously
Scalable Enterprise Influencer CampaignsA tech-enabled agency built for high-volume campaigns, coordinating hundreds of creators simultaneously with end-to-end logistics, content rights management, and product seeding.Clients: Google, Ulta Beauty, Converse, AmazonVisit Obviously →