Ethical AI Promise

SushiSend uses machine learning to fight abuse, improve deliverability, and make the product better over time. Because that work involves data you entrust to us, we want to be clear and specific about what we do — and what we will never do.

What we use data for

• Abuse prevention: classifying content for phishing, malware, adult content, and CSAM. This is non-negotiable to operate a reputable sending platform.
• Deliverability modeling: predicting bounce and complaint risk, recommending domain warmup schedules, and catching reputation drift early.
• Product quality: measuring API errors, UX friction, and feature adoption so we can prioritize what to improve next.
• Aggregated analytics: anonymized, workspace-agnostic trends used in public benchmark reports.

What we will never do

• Train general-purpose foundation models on your email bodies, contact lists, or recipient data.
• Share your content with third-party AI vendors for their model training.
• Use any one customer's data as training input for another customer's sending.
• Sell or lease your data to data brokers or advertisers.

Your controls

• Workspace-level opt-out: toggle "Exclude my workspace from training" in Settings → Data. Applies to all data from that workspace going forward.
• Per-message opt-out: set x-sushisend-training: opt-out on the API request to exclude a specific send.
• Data deletion: request deletion at any time — we remove your data from training sets within 30 days.
• Transparency: we publish a quarterly report describing which models we run, what they're trained on, and material changes to this promise.

Human review

Abuse classification is reviewed by humans before any account-impacting action is taken. We operate under a two-person rule for permanent suspensions and publish an annual appeals report covering resolution times, overturn rates, and reviewer demographics.

Independent audits

Starting in our first full calendar year of operation we will commission an annual independent audit of this promise and publish the results.

Questions

Email ethics@sushisend.com. We respond to every message, publicly when useful.