AI Chatbots and Data Collection: What the Surfshark Ranking Means for ChatGPT Alternatives

AI chatbots are increasingly used for work, study, and everyday tasks—but they differ significantly in what data they collect and how that data may be linked back to you. A recent Surfshark ranking highlights these differences by comparing popular chatbots based on their data-collection footprint. This article explains what such a ranking typically measures, why it matters, and how to use it when evaluating ChatGPT alternatives.

What a “data collected” ranking usually captures

When an analysis ranks AI chatbots by how much data they collect, it’s generally summarizing the categories of information an app or service says it can collect (often from app store privacy labels, product documentation, and policy disclosures). While the exact methodology varies, the most common data categories include:

Identifiers: account IDs, email address, device identifiers, advertising IDs.
Usage data: how you interact with the app (features used, session length, clicks, in-app events).
Diagnostics: crash logs, performance metrics, error reports.
Content you provide: prompts, uploaded files, chat history, feedback.
Purchasing information: subscription status, transaction metadata (often processed via payment providers).
Location data: coarse or precise location (not always collected, but sometimes available depending on platform permissions).

Rankings typically count how many of these categories appear in disclosures, and sometimes distinguish between data used for core functionality (e.g., storing your chat history) versus data used for analytics or advertising.

Why the amount of data collected isn’t the whole story

It’s tempting to assume “less collected = safer,” but a meaningful privacy evaluation needs a few additional questions:

Purpose limitation: Is the data used only to provide the service, or also for marketing, analytics, and model improvement?
Linkability: Is collected data tied to a real identity (account email) or kept separate/aggregated?
Retention: How long is chat content and telemetry stored? Are there deletion controls?
Training use: Are user conversations used to train models by default, or only with opt-in?
Access controls and sharing: Is data shared with third parties (analytics SDKs, ad networks) or restricted to essential processors?

In other words, two chatbots might “collect” a similar set of data types, yet differ drastically in how responsibly that data is handled.

How to use the Surfshark-style ranking when choosing ChatGPT alternatives

If you’re comparing ChatGPT alternatives (or deciding whether to use multiple tools), a data-collection ranking is a helpful starting filter. Use it to narrow candidates, then validate with a quick checklist:

Check account requirements: Can you use the chatbot without creating an account? If not, what minimum identifiers are required?
Review training controls: Look for clear settings to opt out of using chats for training, or enterprise/workspace modes that exclude training by default.
Inspect chat history behavior: Is history stored by default? Can you disable history, set auto-delete, or export/delete data easily?
Look for third-party tracking: Mobile apps may include analytics frameworks. Prefer services that disclose them clearly and offer opt-outs.
Choose “least sensitive input” habits: Even with strong policies, avoid entering passwords, private keys, personal medical/legal identifiers, or confidential business data unless you have a contractual/enterprise setup.

Common trade-offs: privacy vs. convenience

Privacy-friendly choices can come with practical compromises:

Personalization: Less retained history can reduce continuity across sessions.
Safety and abuse prevention: Some data collection supports fraud detection and rate-limiting.
Feature depth: Integrations, voice features, and multi-device sync often require more identifiers and telemetry.

The goal is not “zero data,” but appropriate data: collect what’s necessary, minimize what’s optional, and provide transparent controls.

Practical recommendations for privacy-conscious AI use

Use separate accounts: Keep AI tools off your primary email when possible; consider dedicated addresses.
Disable chat history/training where available: Prefer explicit opt-out toggles and verify they apply to your device and account.
Limit permissions: On mobile, deny location, contacts, microphone, and photos unless needed.
Segment workloads: Use one chatbot for general brainstorming and another (with stricter controls) for sensitive professional work.
Read the “what we collect” section: A quick scan of the privacy policy often reveals whether data is used for ads, shared with partners, or retained long-term.

Bottom line