Terms and conditions

Thank you for using Safety First AI. These terms explain what you can expect from us, and what we expect from you, when you use our app. We’ve tried to keep them short and in plain English.

Who we are

Safety First AI is operated by People First Ventures Pty Ltd (ACN 701 204 944), a company registered in Australia. When these terms say “we”, “us” or “our”, they mean People First Ventures Pty Ltd.

“The app” means the Safety First AI web app at app.safety-first.ai, and this website. “You” means the person using the app.

By creating an account or using the app, you agree to these terms. This is also true when you sign up with a social account instead of an email address and password. If you don’t agree, please don’t use the app.

You can reach us at [email protected].

Changes to these terms

We keep improving the app, so these terms will change over time. The date at the top of this page shows when they last changed. If a change is significant, we’ll let you know by email or in the app before it applies. If you don’t agree with a change, you can stop using the app and cancel your plan.

Your account

• You must be a human. Accounts created by bots or other automated methods aren’t allowed.
• You must be at least 18 years old.
• Each account is for one person. Don’t share your sign-in with anyone. Invite them to your organization instead.
• You’re responsible for keeping your password safe. We recommend turning on two-factor authentication in your account settings. We aren’t responsible for any loss caused by someone else using your password.
• You’re responsible for everything that happens under your account.

Every account belongs to at least one organization: a shared space where a team uses one plan and shares its AI credits. An organization has one owner, and can also have admins and members. The owner and admins manage the plan, payments and members. They’re responsible for the people they invite, and for making sure those people follow these terms.

AI answers can be wrong

Every AI model sometimes states something false as if it were fact, such as a statistic, a quote, a court case or a research paper that doesn’t exist. Asking several AI models the same question in Multi-AI Chat makes these mistakes easier to spot, but it can’t catch all of them, because several models can agree on something wrong. So:

• Check important facts against the original source before you rely on them.
• Answers from the AI aren’t legal, financial, medical, tax or other professional advice, and they don’t replace your own professional judgment.
• You’re responsible for how you use the AI’s answers, including anything you send to your clients or publish.

Personal details and placeholders

Document Analysis and Report Writing look for personal details in the text you paste, such as names, dates, addresses, phone numbers, email addresses and organizations. They replace each one with a placeholder, such as “Person 1”, before your text goes to the AI. To find these details, your text passes through a detection service that we run ourselves, which doesn’t keep a copy of it.

The detection is automatic. It can miss personal details, and it can mark text that isn’t personal. That’s why you check the placeholders before anything is sent, and you can add or remove any of them. You’re responsible for that check.

Your text goes to the AI exactly as you typed it, without placeholders, when you:

• use Multi-AI Chat
• type a message in the chat of Document Analysis or Report Writing
• choose “Skip and go straight to AI” in Document Analysis or Report Writing

Only send text you have the right to share. If you have a duty to keep your clients’ information confidential, you’re responsible for making sure the way you use the app meets that duty.

AI companies

When you choose GPT, Claude or Gemini, what you send goes to the company that runs that model: OpenAI, Anthropic or Google. Each of these companies handles it under its own terms for businesses that use its models, which we don’t control. When you use their models, you also agree to follow their rules: OpenAI’s usage policies, Anthropic’s Usage Policy and Google’s Generative AI Prohibited Use Policy.

What you may not do

You agree not to:

• break any law, or help someone else break one
• use the AI in the app for anything illegal or harmful
• infringe anyone’s rights, including the rights of other users and our own
• abuse, harass, threaten or impersonate anyone, including our staff
• get around AI credit limits, seat limits, the limit of one free trial per person, or any other limit on paid features
• use bots, scripts or other automated tools to send requests to the app or copy data out of it, except through features we offer for that purpose
• overload or disrupt the app, or try to break into it or test its security without our permission (if you find a security problem, please tell us at [email protected])
• reverse engineer the app, or otherwise try to find out its source code or how it works inside
• resell access to the app

Free trial, plans and AI credits

AI credits pay for the work the AI models do. Each request to an AI model uses credits, and longer texts and answers use more.

Free trial. Every new account starts with a free 7-day trial that includes some AI credits. You don’t need a credit card. Each person can have one free trial. When the trial ends, the AI tools stop working until you choose a paid plan.

Paid plans. Plans are paid monthly or yearly, in advance, and renew automatically at the end of each billing period until you cancel. Each plan includes a number of seats, which is how many people can be members of your organization, and AI credits for every month, yearly plans included. Plan credits you don’t use by the end of a month expire, and don’t carry over to the next month. Prices, credits and seats for each plan are on our pricing page.

Top-ups. While you have a paid plan, an admin can buy top-up packs of extra AI credits. Each pack expires after the time shown when you buy it. Your plan’s credits are used first, then top-ups.

When credits run out, AI requests stop until your plan’s credits renew or you buy a top-up. We never charge you for extra use automatically.

Changing plans. If you move to a higher plan, or from monthly to yearly billing, the change happens right away, and you pay the difference for the rest of the current billing period. If you move to a lower plan, or from yearly to monthly billing, the change happens at the start of your next billing period, and you aren’t charged for the new plan before then. A lower plan may have fewer seats and credits, so you may lose access to some features, or need to remove members. We aren’t responsible for that loss.

Payment

Payments are handled by Stripe. Stripe’s Link service is the seller of every purchase (the “merchant of record”), so your receipt and card statement show it as sold through Link. Stripe also works out and collects any GST, VAT or sales tax that applies where you live.

Prices are in US dollars. If you pay in another currency, your bank may charge a fee to convert it.

For questions about a payment, you can contact us or Link’s support team.

We may change our prices. If the price you pay changes, we’ll email your organization’s owner at least 30 days before. The new price applies from your first renewal after those 30 days, and you can cancel before then.

Refunds

If you aren’t happy with the app within 30 days of your first payment for a paid plan, email us at [email protected], and we’ll give you a full refund. No questions asked.

After those 30 days, payments generally aren’t refundable. But things change and problems happen, so if you think a refund is fair, email us. We’ll consider each request on its own merits.

These rules don’t limit your rights under the Australian Consumer Law.

Canceling and deleting your data

You’re responsible for canceling your plan. An admin of your organization can cancel it at any time in the billing settings. An email asking us to cancel doesn’t cancel your plan by itself, but if you have trouble canceling, email us and we’ll help.

When you cancel, you keep access until the end of the billing period you’ve paid for, and you won’t be charged again. Apart from the 30-day refund described under “Refunds”, we don’t refund the unused part of that period.

Where your work is stored:

• In Document Analysis and Report Writing, your document and your chat with the AI are kept only in your web browser, on your own device. We don’t store them. Clearing this website’s data in your browser deletes them.
• In Multi-AI Chat, a chat you don’t save is gone when you close or reload the browser tab. A chat you save is stored on our servers until you delete it.

You can delete a saved chat, your account or an organization at any time. Deleting an organization also cancels its plan right away, without a refund for the rest of the billing period. When you delete something, we remove it from our database right away, and we can’t bring it back. Copies in our backups are removed as those backups expire.

When we may suspend or close an account

We may suspend or close an account if:

• you break these terms, especially the rules under “What you may not do”
• you don’t pay what you owe
• the law requires it, or keeping the account open would put other users, us or the app at serious risk

Changes to the app

We keep improving the app, so its features may change or be removed. We may also offer beta features for you to try early. These may not work perfectly, and may change or be removed without notice.

Your content

You own the text you send to the app and the answers the AI gives you, as far as the law and the AI company’s terms allow. We don’t claim any rights to them.

You give us permission to store your content, process it and send it to the AI models you choose, only as needed to run the app for you.

We own the app, including its software, design, name and logo. You may not copy or reuse any part of it without our written permission.

Security, uptime and who sees your data

We take security seriously. Connections to the app are encrypted, saved chats are encrypted in our database, and we keep backups. But no service is perfectly secure, and we can’t guarantee that yours will never be affected.

We work to keep the app available at all times, but we can’t guarantee that it will be. The app is provided “as is” and “as available”.

People at Safety First AI look at your data only:

• to help with a support request, after asking for your permission
• to fix an error that stops the app from working, looking at as little of your data as possible
• to investigate a security problem or a report of abuse
• when the law requires it

Our privacy policy explains how we handle personal information.

Complaints

The best way to tell us about a problem is to email [email protected]. We aim to reply within 1–2 business days, and to fix the problem or give you a clear update within 10–15 business days.

If you’re an Australian customer, and you believe you have a claim under your consumer rights that we can’t resolve with you, you can contact the Fair Trading or Consumer Affairs office in your state or territory.

Liability

As far as the law allows:

• We aren’t liable for any indirect loss arising from your use of the app, such as lost profits, lost data, harm to your reputation or a claim made against you by a client. This includes losses from decisions made using the AI’s answers, and from personal details that the detection missed.
• Our total liability to you, for all claims together, is limited to the amount you paid us in the 12 months before the claim.

If you’re a consumer under the Australian Consumer Law, the app comes with guarantees that can’t be excluded, and nothing in these terms limits your rights under that law. Where the law lets us limit our liability for failing to meet one of those guarantees, our liability is limited to supplying the service again, or paying the cost of having it supplied again.

The law that applies

These terms are governed by the laws of Victoria, Australia, and the courts of Victoria can decide any dispute about them.

If a court decides that part of these terms can’t be enforced, the rest still applies. If we don’t enforce part of these terms at some point, we can still enforce it later.

Questions

If you have a question about these terms, email us at [email protected].

Adapted from the Basecamp open-source policies / CC BY 4.0