Legal & Privacy Center

Effective date: 6 September 2026

These Terms of Use (“Terms”) govern your use of opportunita.org, the Opportunità Workspace and every related service (together the “Platform”), operated by Opportunità (“we”, “us”). By using the Platform you agree to these Terms. By creating an account you also enter into a contract with us on these Terms. If you do not agree, please do not use the Platform.

If you are a consumer in the European Union or European Economic Area, nothing in these Terms limits the rights you have under the mandatory consumer-protection law of the country where you live.

1. A few definitions

• Official Source — the career page or application system of the university or research institute that is hiring, shown as the “source” link on every offer.
• Listing or Offer — our structured record of a vacancy published by an Official Source.
• Workspace — the private, account-based area of the Platform: dashboard, saved opportunities, application tracker, resume builder, documents, AI tools, plan and credits.
• Credits — whole-number units that meter the AI tools. They have no cash value.
• AI Tools — the features described on our AI Transparency page that use a machine-learning model to produce a report for you.
• Your Content — everything you enter or upload: profile, preferences, resume content, documents, notes and the inputs you give to AI Tools.

2. What Opportunità is — and is not

Opportunità is a discovery layer. We collect publicly advertised academic positions from Official Sources, put them into comparable fields, and give you tools to search, save, track and prepare. We are not a university, employer, recruitment agency, placement service or application portal. We do not receive, forward, review or decide on applications. We are not affiliated with, endorsed by or acting for any institution whose vacancies we list.

Every application is made on the Official Source. Before you act on a listing — a deadline, a salary, an eligibility rule, a contact address — verify it there. Our Source Disclaimer explains how listings are produced and where they can differ from the original.

3. Eligibility and your account

• You must be at least 16 years old. The Platform is intended for people preparing for doctoral and post-doctoral positions and for those who support them.
• You may hold one personal account. Provide a working email address and keep your details accurate.
• Keep your password confidential. We recommend turning on two-factor authentication in Settings → Security, and reviewing the Sessions page to end any session you do not recognise. You are responsible for activity under your account until you tell us it has been compromised.
• You may sign in with Google or Apple where offered. Your relationship with those providers is governed by their own terms.
• We may refuse, suspend or close an account that breaches these Terms or is used for abuse or fraud. We will tell you why unless the law prevents it.

4. Free use

Browsing, searching and reading listings, universities, countries, guides and this Legal Center requires no account and is free. A free account (“Basic”) adds the Workspace: saving opportunities, the application tracker, the resume builder with PDF export, the rule-based Resume Check, document storage, a copy of your data on request and a small monthly allowance of Credits. We may change what the free tier includes, but we will never remove access to data you have already stored without first giving you the chance to export it.

5. Paid plans, credits and payment

5.1 Plans and prices

Paid plans give you a larger monthly allowance of Credits and, depending on the plan, let a capped number of unused Credits carry over into the next month. The plans, prices, billing intervals, credit allowances, carry-over caps and the credit cost of each AI action are shown on the Pricing page and in your Plan & Credits dashboard at the time you buy, and those figures form part of your contract. Prices are in euro; any applicable VAT is stated at checkout.

5.2 How payment works

Payments are handled by Stripe on Stripe-hosted checkout and billing pages. We never see your card number. A plan starts when Stripe confirms payment. Monthly plans renew every month and “Season” plans every three months, automatically, until you cancel. You can change or cancel at any time from Plan & Credits → Manage billing. Your plan stays active until the end of the period you have paid for and does not renew afterwards.

5.3 Your right to cancel within 14 days (EU/EEA consumers)

You may withdraw from a paid plan within 14 days of purchase without giving a reason. Because Credits are available immediately, completing checkout is your request that we start providing the service during those 14 days. If you withdraw, we refund the price minus a proportionate amount for the part of the period already used and for Credits you have already spent, calculated from the plan’s price and the published credit costs. To withdraw, use the Contact form and tell us the account email and the plan; you may also use the standard EU withdrawal form. The right of withdrawal does not apply to renewals.

5.4 Refunds after the 14 days

Apart from the right above, we do not refund unused time or unused Credits when you cancel, unless the service was unavailable for a significant part of a billing period because of us, or the law requires a refund. If we close your account without cause during a paid period, we refund the unused part.

5.5 Credits

• Credits are spent only when an AI action completes and passes our checks. A failed, interrupted or rejected run never costs a Credit. Running the same action again costs Credits again.
• Monthly Credits refill on your plan’s anniversary date; unused Credits above the carry-over cap expire. Credits expire when your plan ends, unless we granted them with a later expiry.
• Credits are personal, cannot be transferred or exchanged for money, and are not a stored-value product.
• We may change the credit cost of an action or a plan’s allowance with at least 30 days’ notice on the Pricing page and in the Workspace; changes apply from your next billing period.

5.6 Failed payments

If a renewal payment fails, Stripe retries it. If it keeps failing, your plan returns to the free tier at the end of the paid period. Nothing you stored is deleted.

6. AI Tools

The AI Tools are optional features you choose to run on your own resume and on the public text of a listing. They produce suggestions, scores and commentary generated by a machine-learning model. They can be wrong, incomplete or biased. They are not career, legal, immigration or eligibility advice; they do not know the institution’s real selection criteria; and no institution sees them. You are responsible for checking anything an AI Tool tells you before you rely on it, and for every decision you take based on it. The AI Tools never apply on your behalf, contact anyone, write a cover letter or change your resume without your action. Each run asks you to confirm these points. Details, limitations and provider information are on the AI Transparency page, which forms part of these Terms.

7. Your Content

• You keep all rights in Your Content. You give us permission to store, process, back up and show Your Content to you, and to send the parts you choose to our AI provider when you run an AI Tool, solely to provide the Platform. That permission ends when the content is deleted, subject to backups that expire in the normal course.
• Upload only what you have the right to upload. Do not upload documents about other people, confidential material belonging to an employer or institution, or content you are not allowed to share.
• Please do not enter health, religious, ethnic, political or similar sensitive information, or identity-document numbers, anywhere in the Workspace. Our forms do not ask for them and the AI Tools do not need them.
• We do not read Your Content except to run the service, investigate abuse or a security incident, or comply with the law. We do not use Your Content to train AI models and we never sell it.

8. Acceptable use

You agree not to:

• copy, scrape, crawl, mirror or bulk-download the catalogue or any substantial part of it, by hand or by automated means, beyond ordinary personal browsing and the read-only API described in section 9;
• circumvent, disable or probe access controls, rate limits, the consent panel, the outbound gateway or any security feature;
• present Opportunità as an Official Source, or present a listing as if it were your own or an institution’s official communication;
• use the Platform to send unsolicited messages to contact persons named in listings, or to collect their addresses;
• create accounts automatically, share an account, or impersonate another person or institution;
• upload malware, or content that is unlawful, defamatory or infringing;
• interfere with the operation of the Platform or place an unreasonable load on it;
• resell, sublicense or commercially exploit the Platform or the AI Tools without our written agreement.

Automated access by AI crawlers and training bots is governed by our robots.txt; where it disallows access, access is not permitted.

9. Our content and intellectual property

9.1 The catalogue and the Platform

The selection, arrangement, normalisation and presentation of listings, the search and matching logic, the software, design, texts, graphics and the Opportunità name and logo belong to us or our licensors and are protected by copyright, database rights and trademark law. You may use the Platform for your personal, non-commercial purpose of finding and preparing for academic positions, and you may share links to any public page.

9.2 Content that belongs to institutions

Vacancy texts, institution names, logos and marks remain the property of the institutions that published them. We reproduce the factual details and limited excerpts needed to describe a vacancy and refer you to it, always with a link to the Official Source. Institutions can ask for corrections, exclusions or removal as described in the Source Disclaimer.

9.3 Public read-only API and feeds

Our public read-only endpoints and sitemaps exist for indexing, research and light integration. You may use them at a reasonable request rate, without circumventing caching, and only if every listing you display links to its Official Source and credits “via Opportunità” with a link to us. We may throttle, change or withdraw them at any time.

9.4 AI reports

We claim no rights in the reports the AI Tools generate for you. You may keep, print and reuse them for your own purposes.

10. Official Sources and other websites

When you leave the Platform through an “Apply” or “Source” link you pass through a short gateway page that names the destination and then continue to a site we do not control. Those sites have their own terms and privacy policies. We do not guarantee that a destination page still exists or still says what our listing says, and we receive no information about anything you do there.

11. Availability and changes

We aim for continuous availability but do not guarantee it. We may change, add or remove features, and we may label some features “beta” or “experimental”; those may change or disappear without notice. If we ever intend to discontinue the Workspace as a whole, we will give at least 30 days’ notice and keep data export available during that time.

12. Ending the relationship

You may stop using the Platform at any time and ask for your account to be deleted as described in the Privacy Policy. We may suspend or end your access with notice if you materially breach these Terms, and immediately where the breach creates a security, legal or abuse risk. Sections 6, 7, 9, 13 and 14 continue to apply afterwards.

13. Disclaimers

Listings are produced automatically from Official Sources that change without notice. We do not warrant that any listing is accurate, complete, current or still open, that a deadline is correct, that a salary figure applies to you, or that you are eligible for a position. AI outputs are provided as they are. To the extent the law allows, we exclude all implied warranties. Nothing in this section removes our duty to deliver a paid service with reasonable care and skill, or any right you have under your local consumer law.

14. Liability

We are fully liable for damage we cause intentionally or through gross negligence, for death or personal injury, and wherever mandatory law says so. In all other cases, our total liability to you for all claims arising from the Platform in any 12-month period is limited to the greater of the fees you paid us in that period and €50. We are not liable for a missed deadline, an unsuccessful application, a decision you took in reliance on a listing or an AI report, lost profit or opportunity, or indirect loss. If you are a consumer, these limits apply only as far as your national law allows.

15. Law and disputes

These Terms are governed by the law of the country in which Opportunità is established. If you are a consumer living in another EU/EEA country you also keep the protection of the mandatory rules of your own country and may bring proceedings before its courts. The European Commission runs an online dispute-resolution platform at ec.europa.eu/consumers/odr. Please contact us first — most problems are solved quickly through the Contact form.

16. Changes to these Terms

We may update these Terms to reflect changes in the Platform or the law. For material changes affecting members we give at least 30 days’ notice in the Workspace or by email. If you do not accept a change you may close your account before it takes effect; continuing to use the Platform afterwards means you accept the new Terms. The current version is always at this address with its effective date.

17. General

If a provision of these Terms is invalid the rest remain in force. Our not enforcing a provision is not a waiver. You may not transfer your account or these Terms; we may transfer them to a successor operator who takes over our obligations, with notice to you. These Terms, the Privacy Policy, the Cookie Policy, the Source Disclaimer, the AI Transparency page and the Pricing page together form the entire agreement between us.

18. Contact

Use the Report Issue / Contact form on this page or our Contact page. You receive a reference code and we reply by email.

Effective date: 6 September 2026

This Source Disclaimer explains where every listing on Opportunità comes from, what we do to it before you see it, where it can differ from the original, and how to get something corrected or removed. It applies to every offer page, list, count, badge, salary card and summary on the Platform.

1. Official sources only

Each listing is collected from the official career page or application system of the institution that is hiring — a university, university medical centre, research institute or a national research organisation’s own portal. We do not take listings from job boards, aggregators or third-party recruitment sites, and we do not accept paid placements. The exact page a listing came from is shown on every offer as the source link; the “Open official source” and “Apply” buttons take you there.

Our collection software identifies itself to institutions as an Opportunità fetcher, respects each institution’s crawling rules, requests pages at a modest pace and re-checks every source on a regular schedule, usually daily. Where an institution publishes vacancies in more than one language or on more than one portal, we collect each version and merge duplicates.

2. What we copy and how we transform it

From each vacancy we extract the details that let you compare positions: title, institution, department or faculty, location, position type, discipline and research area, contract type, duration, hours and FTE, salary text and, where the institution states them, salary figures, the application deadline, the publication date, the contact person or supervisor named by the institution, the application link, and the description as published. We then normalise these into common fields so that listings from hundreds of sites can be filtered side by side:

• Dates are converted to one calendar format and shown in your interface language. Where the source gives only a date, we treat the deadline as the end of that day in the institution’s local time; always confirm the exact hour on the Official Source.
• Position types (PhD, PostDoc, Researcher, Lecturer and others) are assigned by rules from the vacancy text and title. The institution’s own wording prevails if they differ.
• Salary text is kept as published. Structured salary ranges are stored only when the source states figures; we do not convert currencies, apply pay steps or estimate an individual salary.
• Languages. Vacancies are stored in the language the institution published. Our interface exists in several languages, but the vacancy text itself is not translated. Where a posting is in a local language we add a notice.
• Descriptions are cleaned of navigation and formatting, and long texts may be shortened. The complete text is always on the Official Source.

3. AI-assisted enrichment and summaries

Some fields are completed with the help of a language model that reads the vacancy text and proposes structured values, for example a deadline that appears only in prose, a salary range or a department name. Proposals for critical fields — deadline, salary figures, contact email — are never written to a listing automatically: a person reviews them, or they are applied only when the model is very confident and the value is consistent with the source. The title, source link, application link, description, position type, institution and country are never changed by AI.

On some offer pages you will see an “Offer Summary” card marked as generated by AI. It is a short English paraphrase of the stored vacancy text, produced with instructions not to invent requirements, benefits, salary, visa, deadline or funding details that are not in the source. It can still omit or misstate something. It is hidden automatically when the source text changes, until a fresh summary is produced, and you can switch all AI explanations off with the “AI Explanations” control. Our AI Transparency page describes these systems in full.

4. Reference information shown next to listings

• Salary scheme cards show approximate gross ranges from published collective agreements and pay scales for the country and scheme a vacancy refers to. They are reference data, not the salary you would receive; a range marked “pending verification” has not yet been checked against a current source.
• Ranking signals (QS World University Rankings, QS by Subject) are shown with a link to the ranking publisher and the edition used. Rankings are the publisher’s opinion, not ours.
• Recognised-sponsor badges (for example the Dutch IND public register) mean that the institution appeared on that public register when we last imported it. They are not a promise that any particular position comes with visa sponsorship.
• Institution profiles (type, country, city, website, description) are compiled from the institution’s own site and public registers.

5. Freshness, expiry and “no deadline published”

We re-check sources regularly, but institutions change or withdraw postings without notice. A listing may therefore be temporarily outdated, may show a deadline that has since been extended, or may stay open on our site for a short time after the institution closed it. Listings whose deadline has passed are marked expired, are kept out of search engines, and appear only where we say so, for example in an institution’s recent history. Where the institution published no deadline we say “no deadline published” rather than inventing one; such positions are often filled as soon as a suitable candidate is found, so apply early. Counts on our pages (“N open positions”) are computed from our catalogue at that moment and are not an official statistic.

6. Independence and no affiliation

Opportunità is independent of every institution it lists. Institution names, logos and marks belong to their owners and are used only to identify the source of a vacancy. Listing a position does not mean the institution endorses Opportunità, and nothing on our site is an official communication of any institution. We receive no fee from institutions for listing vacancies.

7. Applications happen on the Official Source

We do not host application forms, do not collect applications and do not pass anything to institutions. When you click “Apply” we show a short gateway page naming the destination and then open the Official Source in a new tab. Anything you do there is between you and the institution. What you record in your own application tracker stays private to your account.

8. People named in vacancies

Institutions often name a supervisor or contact person and publish a work email address so applicants can ask questions. We show those details as published, with the email address protected against automated harvesting, for as long as the vacancy is open. Please use them only for genuine questions about the position. If you are the person named and would rather not appear, or if the details are wrong, use the Data Removal form; we act within five working days. The Privacy Policy, section 11, describes this in detail.

9. Known limits of automated collection

Despite checks and review, the following can happen: a deadline extracted from prose is a day off; a salary figure belongs to a different scale step than the one that would apply to you; a vacancy is assigned a broader or narrower position type than the institution intended; two postings for the same position are merged or, less often, shown twice; a department name is missing; a source page has moved and the link returns an error until our next check. If you spot any of these, please tell us through Report Issue — every report is read and most corrections are applied the same day.

10. Corrections, removal and requests from institutions

• Anyone may report an incorrect, outdated or broken listing via Report Issue. Give the offer URL and, if possible, the Official Source URL.
• Institutions may ask us to correct a listing, to change how we collect from their site (frequency, scope, categories), to exclude specific pages, or to stop collecting from them entirely. Write from an institutional email address through the Data Removal form and choose “Source update”, “Data removal” or “Copyright concern”. We confirm receipt with a reference code and act within five working days; removing a listing is usually immediate.
• Rights holders who believe we reproduce more of a text than is needed to describe and refer to a vacancy may use the same form; we remove or shorten the content while we review.

11. Reusing our data

Our public pages, sitemaps and read-only API may be used for indexing, research and light integration at a reasonable rate, provided each listing you display links to its Official Source and credits “via Opportunità” with a link to us. Bulk copying of the catalogue, and use by AI crawlers our robots.txt disallows, is not permitted. See the Terms of Use, section 9.

12. Contact

Questions about sources and data quality: the Report Issue / Contact form on this page.

Effective date: 6 September 2026

This Cookie Policy lists every cookie and every piece of browser storage that opportunita.org and the Opportunità Workspace use: who sets it, why, for how long, and whether we need your consent. It complements our Privacy Policy.

1. What cookies and browser storage are

A cookie is a small text record a website asks your browser to keep and send back with later requests. Browser storage (localStorage) works similarly but stays in your browser and is not sent to us automatically. Both let a site remember something between visits. We use a deliberately small set, almost all first-party, and never for advertising, cross-site tracking or fingerprinting.

2. The three categories

• Necessary — needed for the site, your sign-in and our security protections to work. Set without consent.
• Functional preferences — remember choices you make about how the site looks and behaves. On by default, off with one click, and they store nothing but your own choice.
• Analytics — measurement of how visitors use the site. Currently not active. If we ever switch it on, it will load only after you accept this category.

We do not use marketing or external-media cookies, so the consent panel does not offer those categories.

3. Complete list

3.1 Necessary

3.2 Functional preferences (on by default, can be turned off)

All functional values are your own choices. None identifies you, none is combined with other data, and none is used for measurement.

3.3 Analytics (currently inactive)

4. Other requests your browser makes

For completeness, these services receive a request from your browser on some pages even though they set no cookie on our domain:

• Google Fonts deliver our typefaces on public pages and in the Workspace. Google receives your IP address and browser details for that request, as with any file loaded from another server. Google states that it does not use Fonts requests for advertising or profiling.
• Stripe pages open only when you start a checkout or open your billing portal; Stripe sets its own cookies on its own domain for fraud prevention.
• Google and Apple sign-in pages open only when you choose that sign-in method and follow those providers’ policies.

We embed no video players, maps, social widgets or advertising pixels.

5. Your consent record

When you save a choice in the cookie panel we keep a record so we can show what you agreed to: the policy version, your choice (accept all, reject all or custom), the Analytics and Functional settings, the page you were on, your interface language, which control you used, the time, an irreversible fingerprint of your IP address and browser (not the address itself) and the country your request came from. Records are deleted after 365 days.

6. Changing or withdrawing your choice

• Open Cookie Preferences from the link in the footer of every page, from the floating dock at the bottom of the screen on desktop, or from the menu on mobile. Accept all, reject all, or set each category.
• Rejecting Analytics removes any analytics cookies we control. Rejecting Functional stops us saving new preferences and clears the existing ones from your browser.
• You can also delete cookies and site data in your browser settings. Clearing storage removes your consent record, so the panel appears again on your next visit.
• Browser “Do Not Track” and Global Privacy Control signals: we run no tracking that these would need to switch off, so they change nothing.

7. Changes

When we add or remove a cookie or change a purpose or lifetime, we update this list and raise the policy version, which makes the panel reappear so you can decide again.

8. Contact

Questions about cookies and storage: the Report Issue / Contact form on this page.

Effective date: 6 September 2026

Opportunità uses artificial intelligence in two very different ways, and this page keeps them apart because your rights and responsibilities differ between them:

• Part A — AI we run on the catalogue. Language models help us structure and summarise publicly published vacancies. We choose to run it, we supervise it, and we are responsible for the result. It never touches your account data.
• Part B — AI tools you run on your own data. Inside the Workspace you can ask a model to review your resume, check your fit for a listing or import a CV. You choose to run them, on your data, for your own purposes, and you are responsible for how you use what they produce. We are responsible for running them securely, honestly and only when you ask.

Part C covers our recommendation engine, which is not a language model but is often mistaken for one. Part D explains how we label AI content and where we stand under EU law. Everything here describes the Platform as it is built; where a feature is switched off, we say so.

Part A — AI we run on the catalogue (our responsibility)

A1. Offer enrichment

After a vacancy is collected, a language model may read its text and propose values for structured fields: deadline, salary minimum and maximum, currency and period, department, faculty, research and field area, contract duration, hours per week, FTE, employment type, supervisor name, contact person and contact email. Every proposal is checked against the source text. Only research area and field area are written automatically. Deadline, salary and all other fields are applied only after a person has reviewed them, or when the model is very confident and the value matches the source. A contact email is never written by the model without a person approving it. Title, description, source link, application link, position type, institution and country are never changed by AI. Each run records which model produced which proposal, so a wrong value can be traced and corrected.

A2. Offer summaries

Some offer pages show an “Offer Summary” card. It is a paraphrase of the stored vacancy text, about 80–130 words, always in English, produced with instructions not to invent requirements, benefits, salary, visa, deadline or funding details that are not in the source and to leave out anything unclear. The card is labelled as AI-generated, carries a “confirm in the official posting” note, is hidden automatically when the source text changes until it is regenerated, and disappears entirely when you switch “AI Explanations” off. The summary never replaces the vacancy text; it is a reading aid.

A3. Salary-scheme and reference verification

Reference cards (salary scales, workload and contract explanations) are periodically re-checked by a model against their published source. The model may only propose; a fixed set of rules accepts a change only when confidence is high, a real source is given, the scheme is unchanged and the change is plausible. Values that fail stay marked “pending verification”.

A4. Institution names

To attach ranking and register badges to the right institution, a model suggests alternative names an institution is known by. Suggestions wait for a person to approve them; nothing you see changes on a suggestion alone.

A5. Oversight and limits

• Catalogue AI processes vacancy text and institution data only. It never receives member accounts, profiles, resumes, documents or behaviour.
• Names of supervisors and contact persons that appear in vacancies may be present in the text the model reads. Those names are already public, are used only to fill the “contact” fields of the same vacancy, and follow the review rules above and the removal route in the Source Disclaimer.
• Runs are capped per day and per month, can be paused with a single switch, and every output is stored with its prompt version and model identifier for audit.
• If a listing looks wrong, use Report Issue; a person corrects it and the model’s proposal for that field is discarded.

Part B — AI tools you run on your own data (your responsibility)

B1. The tools

B2. What is sent to the model — and what is not

• Before any request leaves our server we remove your name, email address, phone numbers and links from the text. The model receives resume text and structured content, the listing’s public text, your chosen profile or lens, and the signals from the Resume Check.
• We never send the uploaded file itself — only text extracted from it. We never send your account details, payment data, other documents, notes or browsing history.
• Text is capped in length, and very short or unreadable text is rejected before any credit is reserved.

B3. Where it goes

Requests go from our server to OpenRouter (United States), an AI gateway that forwards them to the language-model provider configured for the Platform at that moment (for example OpenAI). The provider processes the text to produce the response. We rely on the providers’ data-processing terms, we do not allow our requests to be used to train models where the provider offers that control, and we do not train any model ourselves. The specific model can change and is therefore not named in the interface; you may ask us which model produced a stored report. Transfer safeguards are described in the Privacy Policy, section 9.

B4. What we keep

The validated report (scores, findings, text), the version of the consent notice you accepted, the time, technical metadata (token counts, duration, model identifier) and a credit-usage record. Reports live in your Reports Library, are visible only to you, and can be deleted by you at any time. We do not keep the raw model response or your submitted text beyond the run.

B5. What the tools will never do

• Apply to a position, contact an institution or anyone else, or write a cover letter.
• Change your resume, profile or documents. The Resume Import creates a draft you must confirm; the Review and Fit Check only suggest.
• Show a result to an employer, a university or any other user. No institution has access to the Workspace.
• Charge you for a failed run. Credits are reserved before a run and returned if the model fails, returns unusable output, or the result does not pass our checks.
• Guarantee eligibility, an interview, an offer, a visa or a salary.

B6. Your responsibilities

• Verify before you act. Treat every score, gap and suggestion as an opinion from a system that has read a few pages about you. It does not know the selection committee, the institution’s unwritten preferences, your circumstances or the law.
• Do not rely on AI for eligibility, visa, work-permit, funding or legal questions. The tools are instructed not to infer nationality, immigration status or eligibility, and they will tell you to check the Official Source — do it.
• Upload only your own material. Do not run the tools on another person’s CV or on confidential documents you may not share.
• Keep your inputs honest. Suggestions such as “add a metric, if accurate” are placeholders; putting a number in your resume that is not true is your decision and your risk.
• Read the consent notice each time. It states what will be sent and what the run will cost. Confirming it is your instruction to us to make that transfer.
• Expect errors. Language models invent details, misread scanned text, favour certain writing styles and may reflect biases in their training. If the text you uploaded was low quality (for example a poor scan), we say so and ask you to acknowledge it before running.

B7. Our commitments

• Every run is explicit: nothing is sent to a model unless you press the button and confirm the notice. There are no background AI runs on member data.
• Every report states its own limitations, and every fit report says that it is not a guarantee of eligibility or success.
• Outputs are checked against a fixed structure; praise-only or empty results are rejected and never charged.
• Reports are private, deletable, included in your data export, and erased with your account.
• The methods behind the tools are versioned and improved over time; we do not present them as an expert assessment.

B8. Known limitations

• Quality depends on the model configured at the time and may differ between runs.
• Reports are written in English regardless of the language of your resume or the listing.
• Very long resumes and listings are shortened; thin listings produce weaker fit reports, and we say so.
• Running the same tool again on the same inputs costs credits again; the earlier report stays available for free.
• The “industry fit” lens uses general checks rather than a dedicated industry rubric.

Part C — Recommendations and matching (not a language model)

“Opportunities for you” and the recommendations page come from a rule-based scoring engine that compares your preferences and the evidence in your resume (skills, disciplines, topics, methods, languages) with listings. It is not a language model, it sends nothing to an outside provider, and it never receives your name, email, phone, date of birth, nationality or raw resume text. It runs only if you have switched on personalised recommendations in Settings → Privacy; switching it off invalidates every result immediately. Recommendations show plain-language reasons and unknowns, never a score or ranking of you. The engine is labelled experimental and may be off entirely in some periods. The reverse direction — showing candidates to institutions — is disabled, would require your separate talent discovery consent, and would never release contact details.

Part D — Labelling, decisions and the law

• Labelling. AI-generated text on public pages carries a visible label (“AI-generated offer summary”) and a matching label for screen readers. Reports in the Workspace are produced only at your request and are labelled as AI reports with their limitations. Vacancy descriptions and facts shown as coming from the Official Source are not AI-written.
• No automated decisions about you. No AI or rule-based system on Opportunità makes a decision that has legal or similarly significant effects on you. Nothing we compute is shared with institutions or used to rank you against other people.
• EU Artificial Intelligence Act. We provide the tools in Part B to applicants for their own use, not to employers for recruitment or selection, and no employer or institution can access them. We therefore do not operate them as a high-risk employment system. We meet the Act’s transparency duty by labelling AI-generated content as described above. If we ever offered any matching or evaluation output to institutions, we would reassess its classification and obligations before launch and update this page first.
• Data protection. Our legal basis for sending your text to a provider is the contract with you together with your explicit confirmation for each run; the transfer safeguards and retention periods are in the Privacy Policy.

Part E — Turning AI off

• Hide all AI explanations and context cards on offer pages with the “AI Explanations” switch in the floating dock (desktop) or the menu (mobile). Your choice is stored in your browser.
• Simply do not run the Part B tools; nothing in the Workspace depends on them, and your credits stay unused.
• Keep personalised recommendations off (it is off by default) so no matching profile is used.
• Delete any stored report from the Reports Library at any time.

Part F — Reporting a problem with AI output

If a summary misstates a vacancy, a report contains something offensive or clearly wrong, or an AI feature behaved unexpectedly, use Report Issue and choose “Incorrect data” or “Other”, quoting the offer or report. We review every report, correct catalogue data by hand, and use anonymised problem patterns to improve our instructions and checks.

Changes and contact

We update this page whenever we add, remove or materially change an AI feature, change provider, or change what is sent. The effective date at the top identifies the current version. Questions: the Report Issue / Contact form on this page.