AI supplemental terms

1. What SmartAF™ does, in one paragraph

SmartAF™ reads the documents you connect, builds a private, searchable Brain from them, and makes that Brain available inside the AI assistant you already use, such as ChatGPT, Claude, Grok, or Perplexity. When you ask your assistant a question, SmartAF™ finds the relevant passages in your Brain and hands them to your assistant, which writes the answer. SmartAF™ is the memory. Your assistant is the writer. Understanding that split explains most of what follows.

2. Inputs and outputs

"Inputs" are the documents, Google Drive contents, and questions you provide to an AI Feature. "Outputs" are the passages SmartAF™ retrieves and the answers your assistant generates from them.

1. You are responsible for your Inputs, as section 4 of the Terms of Service describes, and for how you use Outputs.
2. Outputs are only as good as the Inputs behind them. If a document in your Brain is out of date or wrong, answers built from it will be too.

3. How SmartAF™ handles your files

1. Read-only. SmartAF™ reads your files where they live. It does not move, edit, rename, or delete anything in your Google Drive, and the access it requests from Google Drive is read-only.
2. Partitioned. Your Brain is separated from every other customer's. No one outside the accounts you authorize can query it, including us, except for support you request or security work we are legally required to do, and then only to the extent needed.
3. Encrypted in transit (TLS) and at rest.
4. Independently reviewed. SmartAF™ is audited by an outside security firm and certified at CASA Tier 2, the review Google requires before an app may connect to Google Drive, as of the date at the top of these terms.
5. As of the date above. These controls describe the service, including features offered as early access, as of the date at the top of these terms. We may change how we deliver them as we improve the service; a material reduction is notified under section 16 of the Terms of Service.
6. Deleted on your schedule. Disconnect Google Drive or delete a Brain at any time. Archived 90 days, then permanently gone, as section 4(d) of the Terms of Service states.

4. Training, stated plainly

We do not use your Inputs or Outputs to train, fine-tune, or improve any AI model, and our processing providers are contractually prohibited from doing so. We never use one customer's Brain to answer another customer's question. We may look at aggregate, de-identified usage measures (how many documents, how many queries, which features are used) to run and improve the service, and nothing in those measures can be tied back to the content of your files.

5. Ownership of inputs and outputs

1. Inputs are yours. Nothing changes about who owns your documents because SmartAF™ read them.
2. Outputs from your own Brain are yours to use for any lawful purpose. Outputs may not be unique; two people asking a similar question of similar material may get similar answers, and we make no claim that an Output is eligible for copyright protection.
3. Outputs from an Expert Brain are split. You own the part of an Output that is unique to you and your situation. Any part that contains, restates, or is based on the Expert's material stays the Expert's, and you receive a non-exclusive, personal license to use it for your own business or personal purposes while you have access to that Brain. You may not republish, resell, teach, or package the Expert's material as your own.
4. The connector, the persona instructions, and the structure are ours. The instructions, indexes, and structure SmartAF™ builds so your assistant can use your files, and the persona instructions behind every Expert Brain, are AutomationEdit's intellectual property, and are licensed to you only while you use the service.

6. Model providers and your AI assistant

1. Your assistant is your relationship. ChatGPT, Claude, Grok, Perplexity, and any other assistant you connect to SmartAF™ are provided by their own companies under their own terms and privacy policies. When you ask a question, SmartAF™ sends the relevant passages from your Brain to that assistant so it can answer. What the assistant's provider does with that conversation is governed by your agreement with them, including any setting you have chosen about training. We recommend you check that setting.
2. Our own processing providers. To index your documents and find the right passages, SmartAF™ uses third-party AI and cloud providers under written agreements that prohibit them from training on your content. We do not name them in these terms so we can change providers without changing your agreement; the current list is available on request at [email protected].
3. By using an AI Feature you authorize us to transmit your Inputs and Outputs to these providers for the sole purpose of providing the service to you.

7. Accuracy and your responsibility to check

AI assistants make mistakes. A Brain reduces them by grounding answers in your own documents, and it does not eliminate them. Outputs may be incomplete, out of date, or wrong, and may present a confident answer to a question your documents do not actually address. You are responsible for reviewing any Output before relying on it, and you should not act on an Output for legal, tax, medical, financial, regulatory, or safety purposes without independent verification. We are not liable for decisions made in reliance on an Output.

8. Expert Brains

An Expert Brain is built from a coach's, teacher's, or expert's own material and made available to their audience through SmartAF™, in a private workspace rather than in our community. Ask Ulka is one.

1. The Expert's content is the Expert's. AutomationEdit built the Brain and runs it; the knowledge in it belongs to the Expert. We do not endorse, verify, or take responsibility for an Expert's advice, methods, or claims.
2. It answers in the Expert's voice, and says it is an AI. An Expert Brain answers in the Expert's first-person voice, under a license from the Expert, and discloses that it is an AI. Its answers are generated by an AI from the Expert's material. They are not a personal communication from the Expert and do not create a coaching, advisory, or professional relationship with them or with us.
3. Access can be gated. An Expert may limit access to enrolled students, a cohort, or a paid audience, and may set how deep the Brain answers. When your enrollment or purchase ends, so does your access, and any locally saved Outputs remain subject to section 5(c).
4. Disclosure. Each Expert Brain carries the Expert's own results disclosure. Read it. Their results, and their students' results, are theirs.
5. No extraction. You may not use an Expert Brain, or any AI Feature, to systematically extract, reconstruct, or reproduce the Expert's material, or to develop, train, or improve any AI or machine-learning model, product, or service. Automated querying, scripted access, and bulk export of Outputs are prohibited.

9. Prohibited uses of AI Features

In addition to section 5 of the Terms of Service, you will not use any AI Feature to:

1. develop, train, or improve any AI or machine-learning model, technology, or service;
2. present an Output as solely human-written where doing so would mislead someone;
3. impersonate a real person, or generate content in an Expert's voice for use outside the Brain, without that person's explicit consent;
4. generate content that is illegal, infringing, defamatory, fraudulent, or deceptive;
5. make decisions in a "high-risk" use under applicable law (employment, credit, housing, medical, or legal determinations about a person) without human review;
6. circumvent plan limits, access gates, or rate limits by any means.

10. Usage limits

Each plan's document limit, seat count, and any query or rate limit are stated on the pricing page or in your order, and may be shown inside the product. We can apply temporary limits without notice to protect the service from abuse or an outage, and we will tell you if a limit we apply affects your account for more than a day.

11. Early access and beta features

Some features, and at times the whole product, are offered as early access, preview, or beta. Those features are provided as is, may be changed, paused, or withdrawn without notice, may contain errors, and should not be relied on for anything important. We may ask for your feedback, which section 10 of the Terms of Service governs. Early access features are not covered by any refund, credit, or service commitment, and section 15(c) of the Terms of Service does not apply to them. If we ask you to keep an early feature confidential, you will. We are not liable for any loss arising from an early access feature, to the fullest extent the law allows.

11A. If an AI Feature is causing you harm

1. Tell us in writing. If you believe an AI Feature is producing output that is causing you or your customers quantifiable harm, including output you believe is unlawfully discriminatory, email [email protected] with a description of the output and the harm. We acknowledge within two business days and work in good faith to correct, limit, or withdraw the feature.
2. Regulators. If either of us receives notice of a regulatory inquiry about an AI Feature, that party tells the other in writing within five business days, and both cooperate in good faith with the response.
3. We use reasonable care in how we design and configure AI Features. We are not responsible for discriminatory or harmful output caused by the documents you supplied, by changes you made, by your use outside these terms, or by drift in a third-party model we do not control.

12. Changes to AI Features

AI Features depend on providers and models we do not control. We may change which models or providers power a feature, or retire a feature, and will give notice under section 15(c) of the Terms of Service where the change removes something you pay for.

13. Disclaimer and liability

AI Features and Outputs are provided as is and as available, without warranty of any kind. Sections 11 through 13 of the Terms of Service apply to AI Features in full, including the limit on our total liability.

14. Changes to these terms

Section 16 of the Terms of Service applies: 14 days' notice by email for changes that affect your rights or your bill. Where these terms are presented as a click-through, continuing past the notice date is acceptance; where you are asked to accept again at login, you can decline and close your account without a further charge.