This Privacy Policy explains how Perceptron AI Labs Inc. ("Perceptron," "we," "our," or "us") collects, uses, discloses, retains, and protects information when you access or use Perceptron, Auta, perceptronai.org, Perceptron APIs, and related services (collectively, the "Services").
By using the Services, you acknowledge this Privacy Policy. If you use the Services on behalf of a company, organization, or other legal entity, you represent that you are authorized to provide information to us and to cause that entity to be bound by this Privacy Policy and our Terms and Conditions.
This Privacy Policy is designed for an AI dataset creation and annotation platform. The Services may process uploaded images, videos, prompts, URLs, annotations, metadata, generated outputs, usage logs, and publicly available internet content requested through Auta or related tools.
1. Scope
This Privacy Policy applies to information we collect through our websites, applications, APIs, demos, documentation, beta features, support channels, events, sales communications, and any service that links to this Privacy Policy.
This Privacy Policy does not apply to third-party websites, third-party datasets, third-party models, third-party integrations, or services that we do not control. Those third parties may have their own privacy policies, licenses, and terms.
This Privacy Policy is incorporated into our Terms and Conditions. Capitalized terms not defined here have the meanings given in the Terms and Conditions.
2. Roles: Controller, Processor, and Service Provider
For account information, website analytics, billing information, marketing information, security logs, product telemetry, and other business information, Perceptron generally acts as a controller or business under applicable privacy laws.
For non-public Customer Content that you upload into a workspace under a written enterprise agreement or Data Processing Addendum, Perceptron may act as a processor or service provider, processing that data on your behalf and under your instructions.
Unless a separate written agreement expressly states otherwise, Perceptron may process Customer Content, Public Source Content, Output, logs, and derived information as described in this Privacy Policy and the Terms, including for service operation, security, abuse prevention, analytics, research, development, testing, training, and improvement of Perceptron Technology.
You are responsible for determining whether your use of the Services makes you a controller, business, processor, service provider, joint controller, or other regulated party, and for satisfying your own privacy, notice, consent, and data-processing obligations.
3. Information We Collect
We may collect the categories of information below, depending on how you use the Services:
- Account and profile information, such as name, email address, organization name, role, login credentials, authentication information, and user preferences.
- Commercial and billing information, such as plan information, invoices, payment status, billing address, purchase history, and payment processor references. We may use third-party payment processors and do not intentionally store full payment card numbers unless expressly stated.
- Customer Content, including images, videos, zip files, URLs, prompts, instructions, annotation schemas, labels, masks, bounding boxes, polygons, OCR results, captions, notes, metadata, datasets, exports, and other materials you or your authorized users provide.
- Public Source Content, including public images, videos, webpages, captions, URLs, license signals, source metadata, thumbnails, and related materials that the Services identify, retrieve, cache, process, annotate, transform, or export in response to your dataset requests or our service operations.
- Output, including annotations, labels, structured data, masks, bounding boxes, generated metadata, model responses, derived features, dataset organization, evaluations, and exports generated by or through the Services.
- Usage and device information, such as IP address, browser type, device identifiers, operating system, pages viewed, referring URLs, API calls, latency, errors, feature usage, model/version identifiers, token or compute usage, timestamps, and diagnostic logs.
- Communications, including support requests, sales inquiries, demo requests, feedback, survey responses, messages, and recordings or transcripts where disclosed.
- Integration information, such as files, metadata, permissions, access tokens, repository references, cloud-storage references, and other information you choose to connect to the Services.
- Cookies and similar technologies, including analytics identifiers, session cookies, local storage, pixels, and related tracking technologies used to operate, secure, measure, and improve the Services.
- Information from third parties, such as authentication providers, analytics providers, enrichment providers, public sources, event partners, affiliates, service providers, and other users in your workspace.
4. Public Internet Dataset Sourcing
Auta and related Services may allow you to describe a dataset and request that the Services source imagery or other materials from publicly available internet locations or third-party repositories.
When this feature is used, we may collect and process Public Source Content, source URLs, crawl or retrieval metadata, license signals, file metadata, deduplication fingerprints, embedding vectors, annotations, quality scores, and other technical information needed to create, evaluate, organize, filter, annotate, and export datasets.
Public availability does not mean that content is free of copyright, privacy, publicity, contractual, license, or other restrictions. We do not represent that Public Source Content is rights-cleared, accurate, current, safe, lawful for your intended use, or suitable for model training, commercial use, publication, or redistribution.
You are responsible for reviewing Public Source Content, source terms, licenses, attribution requirements, privacy implications, rights-clearance needs, and downstream use restrictions before using, training on, distributing, publishing, or commercializing any dataset or Output.
We may remove, block, de-index, filter, quarantine, or stop using Public Source Content or source domains at our discretion or in response to rights-holder requests, privacy requests, source restrictions, law, abuse reports, or our internal policies.
5. How We Use Information
We may use information for the following purposes:
- To provide, operate, maintain, secure, troubleshoot, and support the Services.
- To create, process, annotate, transform, organize, validate, deduplicate, export, and manage datasets.
- To respond to prompts, instructions, API calls, support requests, sales inquiries, and administrative communications.
- To personalize features, remember preferences, manage accounts, authenticate users, and enable collaboration in workspaces.
- To monitor performance, reliability, errors, latency, capacity, quality, abuse, fraud, security incidents, and policy violations.
- To develop, test, train, fine-tune, evaluate, benchmark, debug, and improve Perceptron Technology, including AI models, data pipelines, annotation systems, dataset sourcing systems, ranking systems, filters, safety systems, and product features.
- To create and use aggregated, anonymized, de-identified, statistical, or derived information for any lawful business purpose, including analytics, benchmarking, model evaluation, research, and service improvement.
- To enforce our Terms, this Privacy Policy, the Acceptable Use Policy, the Copyright/DMCA Policy, and other agreements.
- To comply with law, legal process, sanctions, export controls, court orders, regulatory requests, and lawful government requests.
- To protect the rights, property, safety, security, and integrity of Perceptron, users, rights holders, source websites, the public, and third parties.
- To send service notices, security alerts, administrative messages, marketing communications, newsletters, product updates, and event information where permitted by law.
- To evaluate or complete a merger, acquisition, financing, reorganization, bankruptcy, sale of assets, or similar corporate transaction.
6. Broad Service Improvement, AI Training, and Research Rights
Subject to applicable law and any separate written agreement signed by Perceptron, we may use Customer Content, Public Source Content, Output, prompts, instructions, annotations, metadata, logs, feedback, interactions, and derived information to develop, train, fine-tune, test, evaluate, benchmark, monitor, debug, protect, and improve Perceptron Technology and related services.
This may include using information to improve dataset sourcing, object detection, segmentation, tracking, labeling, OCR, captioning, ranking, deduplication, filtering, quality assurance, safety, rights-management workflows, and other AI or data-processing systems.
We may retain and use generalized learnings, model weights, embeddings, statistics, error patterns, quality metrics, evaluations, safety signals, usage patterns, and other derived or de-identified information, even after specific Customer Content is deleted, to the extent permitted by law and our agreements.
Unless a separate written agreement states otherwise, you should not provide sensitive, regulated, confidential, proprietary, or restricted information to the Services if you do not want it processed for the purposes described in this Privacy Policy.
Perceptron does not claim ownership of your Customer Content solely because we process it. However, you grant us the rights necessary to process it as described in this Privacy Policy, the Terms, and any applicable agreement.
8. How We Disclose Information
We may disclose information in the following circumstances:
- To service providers, subprocessors, contractors, consultants, cloud providers, hosting providers, security providers, analytics providers, payment processors, customer support tools, AI/model providers, and other vendors who process information for us or help operate the Services.
- To affiliates and corporate group companies for business, operational, security, support, analytics, and improvement purposes.
- To workspace owners, administrators, collaborators, team members, and authorized users according to workspace settings and permissions.
- To third-party integrations, platforms, storage providers, repositories, APIs, or applications that you connect to or instruct us to use.
- To rights holders, complainants, users, affected parties, or legal representatives when we process copyright, privacy, safety, abuse, or legal requests, including forwarding notices and counter-notices where appropriate.
- To law enforcement, regulators, courts, public authorities, or other third parties where we believe disclosure is required or appropriate to comply with law, enforce agreements, protect rights, prevent harm, or respond to legal process.
- In connection with a merger, acquisition, financing, reorganization, bankruptcy, due diligence, sale of assets, or similar transaction.
- With your consent or at your direction.
- As aggregated, anonymized, de-identified, statistical, or derived information that does not reasonably identify you, for any lawful business purpose.