Skip to content

By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

How I AIHow I AI

Token fraud is here and it’s BAD. Here’s what you need to know.

Today I’m walking you through a problem that has cost me tens of thousands of dollars over the past year and eaten almost all of my engineering energy: token fraud.

Claire Vohost
Nov 16, 202639mWatch on YouTube

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

Great conversations run long. Their best ideas don't have to.

uListen is free on iPhone, iPad, Android, Chrome, and Claude.

See how

At a glance

WHAT IT’S REALLY ABOUT

Token fraud drains AI startups: abuse patterns and practical defenses

  1. Claire Vo describes how token fraud—abuse of an AI product’s paid LLM tokens by non-customers—has consumed most of her engineering effort and cost her tens of thousands of dollars.
  2. She outlines six abuse patterns she has observed, including trial farming, promo code abuse, payment fraud, proxy-based reselling/relaying, MCP-based bypasses, and model probing.
  3. She argues that many effective defenses are business-level tradeoffs, like ending free trials, tightening promo code policies, charging up front with strict billing, and enforcing hard usage caps per account and organization.
  4. She details a technical defense stack spanning AI gateways with model allow-lists, live spend monitoring and caps, payment-fraud tooling, bot-proofing, firewall and rate limits, and tighter controls on custom MCP hosts.
  5. She explains how she uses AI agents to classify suspicious activity, monitor signups and chats hourly, and equip support workflows to handle fraud while preventing support from being used as an attack surface.

IDEAS WORTH REMEMBERING

5 ideas

Token fraud is “stealing inference” through your product’s access to paid LLM tokens.

Claire Vo explains that when an AI app is priced by seats but incurs real per-token API costs, abusers can exploit free access paths (trials, promo codes, early access) to run up the founder’s model bill with non-product-aligned usage.

Token fraud shows up as repeated, creative abuse patterns—not just one-off misuse.

She lists six common patterns she has seen in ChatPRD: free-tier/trial farming, promo code abuse, bad cards/chargebacks, resell-and-relay via proxies, abusing MCPs to bypass controls, and model probing to identify models/cutoffs.

The strongest business controls against token fraud often reduce PLG conversion and pipeline.

She says the most effective levers often hurt legitimate growth: turning off trials, restricting promo codes, charging up front with no billing flexibility, and imposing hard usage caps per account/org.

You have to manage AI cost risk in dollars and time-to-detection, not just request counts.

She recommends technical measures like using an AI gateway to control spend and restrict allowed models, monitoring spend at least hourly/live, and setting alerts, caps, failovers, and automations so exploitation can’t rack up costs quickly.

Payments, identity, and edge controls are core parts of defending LLM spend.

She describes using Stripe Radar aggressively for payment fraud/card testing, plus rules like fingerprinting cards used across multiple trials, and using infrastructure controls like bot proof, firewall country/VPN blocks, IP blocks, and rate limiting.

WORDS WORTH SAVING

5 quotes

I have to tell you, the last year I have been spending more and more of my time, and almost all of our engineering energy, focused on blocking one thing: token fraud.

— Claire Vo

This has cost me tens of thousands of dollars over, I would say, the past year. It has been incredibly painful.

— Claire Vo

Every time I do something that protects us from fraud, it has a negative impact on our real business and our real prospects.

— Claire Vo

I think free trial's over. I'm calling it right now. It's just not worth it.

— Claire Vo

In conclusion, if you sell AI, you sell compute, so act like it.

— Claire Vo

Read the available transcript
Definition of token fraud and seat-based pricing riskFree trial farming and promo code abusePayment fraud, bad cards, and chargebacksReselling/relays via proxies and reverse proxiesMCP abuse and custom MCP host restrictionsAI gateway model allow-lists and spend caps/alertsBot proofing, firewall blocks, rate limiting, and auth-provider controls

Ask ChatGPT about this episode

Get its key ideas, plus a summary of each chapter linked to that point in the video, right in your chat.

  1. Add uListen AI to ChatGPT.

  2. Paste this prompt into a new chat.

    Use uListen AI to give me this episode's key ideas and a chapter-by-chapter summary, with timestamps: https://www.youtube.com/watch?v=ZDpP7JXPlhc

SPEAKERS

  • Claire Vo

    host

    AI podcast host and solo founder of ChatPRD, an AI platform for product managers.

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.