#Social Media, Online Conduct and Misinformation Policy

1.1 The authoritative list. The only accounts, domains and addresses that speak for Pensieve Labs are the following.

1.2 Anything not on that list does not speak for Pensieve. A handle that resembles one above, a domain with an added, altered or transposed character, an email from a free webmail account bearing the brand, and a message on a channel not listed here are not Pensieve, whatever they claim. Pensieve maintains the current list at https://trust.pensievelabs.org and updates it there first.

1.3 How to verify. If you are unsure whether a communication is genuine, do not act on it. Check the sender against 1.1, confirm the account against the list published at https://trust.pensievelabs.org, and where anything is still in doubt, write to info@pensievelabs.org and ask. Pensieve will confirm or deny within two Business Days.

1.4 Recruitment. Pensieve does not recruit through consumer messaging apps, does not ask a candidate for a payment, deposit or fee at any stage, and conducts recruitment only from an address at the [TO BE SUPPLIED] domain. The full position, and what to do if you are approached by someone impersonating Pensieve in a recruitment context, is in the Candidate and Recruitment Privacy Notice (POL-IN-320). Impersonation of Pensieve to defraud a jobseeker is treated under this Policy as seriously as any other impersonation.

1.5 Security contact. A report of a spoofed account, a phishing message or a fraudulent domain using the marks may be sent to info@pensievelabs.org as well as info@pensievelabs.org, and is handled under the Vulnerability Disclosure Policy (POL-GL-059).

2.1 The commitment. Pensieve does not act, under this Policy or otherwise, against any of the following, whoever publishes it and however unwelcome Pensieve finds it.

2.1.1 A negative review, rating or comment.

2.1.2 An honestly held opinion, including a harsh one.

2.1.3 A comparison with another product, favourable or unfavourable to Pensieve.

2.1.4 Journalism, analysis, commentary and criticism.

2.1.5 Academic or research writing about the Platform.

2.1.6 Satire and parody that a reasonable reader would recognise as such.

2.1.7 A truthful account of a person's own experience with Pensieve or with its support.

2.1.8 Security research conducted within the Vulnerability Disclosure Policy (POL-GL-059).

2.1.9 A protected disclosure by a member of personnel under the Whistleblower and Grievance Policy (POL-IN-308).

2.2 Truth is a complete answer. In every market in which Pensieve operates, save one, a true statement is a complete defence to a claim in defamation, and an honestly held opinion is protected. Pensieve will not use this Policy, a contract term, a non-disparagement clause or the threat of legal action to suppress a statement that is true or an opinion that is honestly held. The single exception, the United Arab Emirates, where truth and good faith are not absolute defences, is stated plainly at 11 so that no reader is misled by a blanket assurance.

2.3 No customer is gagged. No Edsol Edtech Pvt. Ltd. customer contract requires a customer, or a customer's staff, to obtain Pensieve's approval before publishing an unfavourable opinion about the Platform, and none prevents a customer from speaking truthfully about its own experience. Reference and publicity run in the other direction only: Pensieve needs the customer's consent to name the customer, not the reverse. See 8 and the Reference and Publicity Consent (ADD-GL-020).

2.4 Disagreement is not a trigger. That Pensieve considers a statement unfair, one-sided, or wrong in its conclusions is not, by itself, a ground for any step in this Policy. Only a false statement of verifiable fact, or one of the specific categories of conduct at 4, is.

Pensieve publishes more ways to complain than to be sued, and it does so first, on purpose. If you have a problem with the Platform or with how Pensieve has treated you, each of these routes reaches a person whose job is to answer you, and each is faster than a post.

Pensieve does not restate those timelines here; each owning document states them and is the single source of truth for them. The point of this clause is simple: a public post is never the only route open to you, and it is rarely the fastest.

This is the closed list of conduct that Pensieve treats as actionable. It is closed on purpose. Anything not on it, and in particular anything at 2, is not actionable under this Policy. Each item is defined so that a reader can test whether a given post falls inside it or outside it.

4.1 Impersonation. Passing off an account, page, message, domain or profile as Edsol Edtech Pvt. Ltd., Pensieve Labs, the Platform or a named member of its personnel, in a way likely to make a reasonable person believe the communication is from Pensieve when it is not.

4.2 Spoofed and lookalike identities. Registering or operating a domain, handle or page that imitates an identifier at 1.1, including by adding, altering or transposing a character, and any paid promotion of such an identity.

4.3 Fabricated and doctored material. Creating, altering or selectively cropping a screenshot, document, invoice, ticket transcript or record so that it misrepresents the Platform or Pensieve, and presenting a screenshot of another product as though it were the Platform.

4.4 False statements of verifiable fact. Publishing, as fact, a statement about the Platform, its security, its availability, its certifications, its pricing, its customers, its personnel or its financial position that is verifiably false. A statement of opinion, a prediction, a comparison and a matter genuinely in dispute are not within this item. On certifications specifically, Pensieve holds none, and the accurate position is published in its Statement of Applicability (STM-GL-010); a false claim that Pensieve is certified is as much a misstatement as a false claim that its security has failed.

4.5 False statements about customer support. A fabricated support ticket or transcript, an invented response time, a claimed outage that did not occur, and a claim of an unresolved complaint that was never in fact raised. A genuine account of a real support experience, however critical, is not within this item and is protected under 2.

4.6 Fake and manipulated reviews. A review, rating, testimonial or reference that is fabricated, incentivised, traded, or written by a person with an undisclosed connection to its subject, in either direction, including a favourable one. The authenticity standard Pensieve holds itself and others to is the Customer Feedback, Reviews and Testimonials Authenticity Statement (STM-GL-035).

4.7 Coordinated inauthentic behaviour. Review brigading, bot amplification, sock-puppet networks and other coordinated activity designed to manufacture the appearance of organic sentiment, for Pensieve or against it.

4.8 Publication of confidential information. Publishing Pensieve confidential information, source code, credentials or a security finding outside the coordinated route in POL-GL-059.

4.9 Publication of a patient record or clinical material. Publishing any patient record, clinical image, identifiable individual or hospital screen captured from the Platform. This is the most serious category and is governed by 6.

4.10 Misuse of the marks. Use of the Pensieve Labs or Pensieve marks contrary to the Trademark and Brand Usage Policy (POL-GL-062), including any use that asserts a certification, accreditation, partnership or approval that does not exist.

4.11 The clinical-decision misstatement. The specific claim that the Platform diagnoses, triages, treats or makes a clinical decision. This is false, and it is not merely a commercial misstatement: it is a regulatory one, because the Platform is not a medical device and the boundary is set out in the Clinical Safety Boundary Statement (DIS-GL-028). Pensieve corrects this claim wherever it appears, including where a well-meaning supporter makes it.

4.12 Synthetic impersonation. AI-generated or otherwise synthetic audio, image or video that depicts a member of Pensieve personnel or the Platform in a manner a reasonable person would take to be real, and any undisclosed synthetic endorsement of Pensieve. Where content of this kind is published in India, it is also subject to the labelling obligations for synthetically generated information introduced into the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 with effect from 20 February 2026.

5.1 The ladder. Pensieve responds to conduct within 4 in ordered steps. It starts at the lowest step that fits, climbs only where a step fails, and halts the moment the statement is corrected.

5.2 The Step 0 test. Before any step is taken, the owner records, in writing, the answer to each of the following, and the matter is not advanced until they are answered.

5.2.1 Is the statement one of fact, or of opinion? Opinion stops here.

5.2.2 If it is fact, is it verifiably false, and what is the evidence that it is false?

5.2.3 Is it one of the categories at 4, or is it protected under 2?

5.2.4 What is the current reach of the content, and would any step Pensieve takes give it more reach than it has now?

A statement that is opinion, or that is true, or that is a genuine matter of dispute, is recorded as such and no further step follows. The decision is kept whether or not Pensieve acts, so that the reason for acting, or for not acting, can be reviewed.

5.3 The decision gate before any legal notice. Pensieve does not send a legal notice, and does not commence proceedings, until a named approver has confirmed each of the following in writing. This gate exists because a legal notice sent against a truthful statement or an honest opinion is both unwinnable and counterproductive, and because the act of demanding removal frequently gives content far more attention than it would otherwise have had.

5.3.1 The statement is one of verifiable fact, not opinion.

5.3.2 The statement is false, and the evidence is attached.

5.3.3 The current reach, the reputational effect, and the risk that the notice itself amplifies the content have been weighed.

5.3.4 A quieter route, a private correction or a factual public reply, would not achieve the same result with less risk.

5.4 Criminal process. Pensieve uses criminal process only where a criminal offence is genuinely made out, which in practice means impersonation or forgery, not a commercial disagreement. Pensieve does not use, or threaten, criminal process as leverage in a commercial dispute. Where a matter also involves a demand from or to a public authority, the Legal and Law Enforcement Request Policy (POL-GL-067) governs.

5.5 No step is skipped. No step on the ladder is skipped without the written approval of the Director, recorded with the reason. The ordinary path is Step 0, then Step 1, and no further. The ladder is a sequence, not a menu.

5.6 Halt on correction. At any step, if the statement is corrected or withdrawn, the process halts and no further step is taken. Correction, not punishment, is the object.

6.1 Publication of any patient record, clinical image, identifiable patient, or hospital screen taken from the Platform is prohibited absolutely, whoever posts it, and is the most serious matter this Policy addresses.

6.2 Where Pensieve becomes aware of such a publication, it is escalated immediately as a security incident under the Incident Response Policy (POL-GL-112), and Pensieve acts to have the content removed as a priority, using the accelerated routes the platform and the applicable law provide.

6.3 Pensieve acts to have the content removed even where the person who posted it is a member of a customer's own staff, because the duty runs to the patient whose record it is, not only to the customer. Pensieve informs the affected customer and coordinates with it, and does so without delaying removal.

6.4 This clause does not displace the customer's own obligations as the party responsible for the record. It is Pensieve's commitment to act on what it sees, at once, whatever the source.

7.1 Personnel are bound in detail by the Personnel Public Communications and Social Media Policy (POL-GL-325), which states who may speak for Pensieve, the disclaimer required on a personal account, the prohibition on posting anything about a hospital, a ward, a screen, a device or a deployment, and the rule that a screenshot of the Platform containing a real record is a data incident to be reported, not deleted quietly.

7.2 This Policy does not restate POL-GL-325. Where a member of personnel is unsure, POL-GL-325 governs their conduct and this Policy governs the company's response.

8.1 What Pensieve asks. A customer, and a partner, is asked not to publish a screenshot or export of the Platform that contains a real patient record, a real clinical image, or an identifiable individual. A sanitised or synthetic screenshot, used to describe the Platform, is welcome.

8.2 Reference and publicity. Pensieve names a customer, quotes a customer, or uses a customer's logo only with that customer's written consent, recorded under the Reference and Publicity Consent (ADD-GL-020). Signing an order is not consent for this purpose.

8.3 The reciprocal commitment. The obligation runs to Pensieve, not from the customer. Pensieve will not name a customer without consent, and a customer remains free to speak about its own experience, including critically, without Pensieve's approval. See 2.3.

9.1 Every action is logged. Each action under this Policy is recorded: the report received, the Step 0 decision, any correction published, any notice sent, any takedown requested, any takedown granted or refused, and any proceedings commenced.

9.2 The count is published. Pensieve publishes, in the Transparency Report (POL-GL-068), the number of assessments made, corrections published, notices sent under each template, takedowns requested, takedowns granted, and proceedings commenced in the period. A policy that claims to act against falsehood, and then publishes no account of when it acted, is not accountable. Pensieve publishes the account.

9.3 Pensieve holds its own statements to the same standard. A company that acts against false statements must first be able to show that its own statements are true. Accordingly:

9.3.1 every complaint Pensieve receives is recorded, with a reference, under the Grievance Redressal Policy (POL-GL-066);

9.3.2 the count of complaints is reported in the Transparency Report (POL-GL-068);

9.3.3 no review, testimonial or reference published by Pensieve is paid for, incentivised, traded or suppressed, under the Customer Feedback, Reviews and Testimonials Authenticity Statement (STM-GL-035); and

9.3.4 it follows that any statement Pensieve makes about customer satisfaction is checkable against a published record, and Pensieve makes no such statement that the record does not support.

9.4 This clause is what makes the rest of the Policy defensible. Pensieve's firmness about the truthfulness of others rests on the same discipline applied to itself.

10.1 A person who receives a correction request, a platform report copied to them, or a notice under this Policy may respond, and Pensieve will consider the response before taking any further step.

10.2 That person may raise a grievance under the Grievance Redressal Policy (POL-GL-066), including a grievance that the notice should not have been sent.

10.3 Where Pensieve concludes that a notice was sent in error, that the statement was in fact true, or that it was an opinion honestly held, Pensieve withdraws the notice in writing, and records the withdrawal in the account it publishes under 9.2.

10.4 Pensieve does not require a person to sign a non-disparagement undertaking, a confidentiality agreement or a release as a condition of resolving a matter under this Policy.

11.1 The table below states, at the level of whether a cause of action exists and is available to Pensieve, the position in each market. It is not legal advice and it is not a legal opinion on any specific statement. Pensieve takes advice in the relevant market before it acts, and the specific instruments and their current thresholds are matters that advice confirms.

11.2 The intermediary timelines in India. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 were amended with effect from 20 February 2026, shortening the removal and grievance timelines to approximately 3 hours for a court or government order, 2 hours for intimate-image and impersonation content, and 7 days for general grievance resolution. [UNVERIFIED: the exact rule-number mapping of the shortened 3-hour, 2-hour and 7-day timelines; three secondary sources agree on the figures but attribute them to slightly different sub-rules. Confirm against the gazette notification of 10 February 2026 before this Policy quotes a specific rule number against a specific timeline.]

11.3 The honest bottom line. In most of Pensieve's markets the fast lever is not a lawsuit. It is a platform's own impersonation, trademark and authenticity process, which acts on impersonation, fabricated material, intellectual-property misuse and fakeness, and which will not remove a genuine critical opinion. Pensieve plans on that reality rather than against it.

11.4 Personal data in a hostile post. Where a post publishes an individual's personal data, for example the doxxing of a member of personnel, the applicable data-protection law may provide a route in addition to the above. In India this is the Digital Personal Data Protection Act, 2023. [UNVERIFIED: which specific obligations under the Digital Personal Data Protection Act, 2023, and which complaint route to the Data Protection Board of India, are in force at the date of reliance; the Act is being commenced in phases. Re-check the phase notifications before relying on this route.]

12.1 Roles.

12.2 Enforcement against personnel. A breach of this Policy or of POL-GL-325 by a member of personnel is a disciplinary matter under the applicable internal policy.

12.3 Review. This Policy is reviewed at least every six months, and on any material change to the law of a market in which Pensieve operates, in particular any change to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 or to the commencement of the Digital Personal Data Protection Act, 2023. The next review is due on 03 February 2027. The shorter-than-usual review cycle reflects how quickly the online-content law in Pensieve's markets is changing.