Over the past 14 years, Instacart has built the infrastructure that powers grocery for millions of households, making shopping faster, easier, and more affordable at scale. AI has become central to that work. It's part of how we help customers discover the right products, match shoppers with the orders, and run a platform efficient enough to make same-day delivery accessible.
As these technologies become more capable and more central to everyday experiences, we want to explain how we use AI at Instacart, the protections we build into its deployment, and how we govern its use.
Today, we’re publishing Instacart's Responsible AI Principles, a plain-language explainer on how we use, deploy, and oversee AI across our entire platform. These principles reflect how Instacart already operates, and we're putting them in writing so anyone can clearly understand where we stand and how we work.
Download: Instacart's Responsible AI Principles
In short, here are the 5 principles that guide how we build and use AI at Instacart:
Transparency. We're transparent about how customers and shoppers interact with AI tools on Instacart.
Here’s how: We disclose AI-generated content when such disclosure is required or relevant, we will clearly indicate when chatbots on Instacart are powered by AI, and we don’t design our AI systems to misrepresent themselves as human. We communicate the reason for shopper deactivations, notify shoppers of the right to appeal, and we don't allow AI systems to make final deactivation decisions.Human oversight. We build AI systems to empower humans, not to replace human oversight.
Here’s how: AI systems can improve consistency and reduce bias, but require ongoing human oversight to help ensure alignment. Our Generative AI experiences are subject to human oversight appropriate to the risk. In addition, where AI contributes to decisions related to shopper deactivations, shoppers will always have the right to human review.Safety. We design AI systems with safeguards to minimize harmful outputs.
Here’s how: Our products are reviewed to mitigate harmful, inappropriate, or unsafe outputs, and we regularly review and update our safeguards to address emerging AI safety risks.Fairness. We design processes to help AI enhance Instacart’s platform while mitigating the risk of unfair outcomes.
Here’s how: We provide shoppers with the opportunity to appeal deactivations, and we review our AI systems to monitor outcomes, and to identify and address potential fairness issues.Governance and accountability. We maintain accountability for how AI is used and deployed across Instacart, through a cross-functional governance framework.
Here’s how: We have executive-level forums and legal and technical working groups. In addition, employees working on these systems can access these groups for guidance and training on these principles.
Why transparency matters
AI is moving fast, and as our use of this important technology grows and changes, we will continue to engage with customers, shoppers, partners, policymakers, and regulators about how we use AI, what we’re building, and what we're learning. Transparency is how we will continue to hold ourselves accountable, and how the people who depend on Instacart can hold us accountable, too.
Our company was built to help everyone access the food they love. AI is one of the most powerful tools we have to deliver on that mission, and we plan to use it well.
