An AI-Written Ph.D. Evaluation: Accountability and Integrity
PYQ Deep Decode
Read the question the way the examiner set it: the demand, the hidden demand, the structure that follows from both, and how the answer should be presented, opened and closed.
1.1 The demand
This is a short case set in Section A, judged through two named values. It is not a question on the merits of artificial intelligence. It asks whether using AI to generate a doctoral evaluation report, followed by unspecified modifications, meets what accountability and integrity require of the professor. Note what the question does not establish: how far he read and verified the thesis, and whether he disclosed the AI use. A strong answer reasons from the facts given and treats the rest conditionally.
| Words in the question | What they ask you to do | What goes wrong if you miss them |
|---|---|---|
| Owing to paucity of time | Treat the reason as a claimed justification to be tested: does shortage of time excuse what follows? | The excuse is either ignored or accepted without examination. |
| a university professor | Hold him to the duties of an academic examiner: expert judgement, care, fairness and confidentiality. | Generic lines on public servants that never reach the examiner’s role. |
| generates a Ph.D. evaluation report using Artificial Intelligence | Examine whether AI merely drafted the report or supplied the substantive academic assessment. The professor must independently verify the evaluation, since it affects the award of a degree. | Treating the AI as a neutral writing aid, like a spell-checker. |
| submits it with some modifications | Ask whether partial editing turns the report into his own judgement, and whether submitting it implies he wrote it. | Assuming that any editing settles the matter. |
| Discuss this from the perspective of accountability and integrity | Examine the act through both lenses separately, with reasons, and reach a view. | One lens only, or a blended list of “ethical issues” with neither value defined. |
1.2 The hidden demand
The question does not ask whether AI should be banned in universities. It tests a narrower and harder idea: responsibility cannot be delegated to a tool. Editing a machine’s draft may improve its language without establishing that the examiner read the thesis and formed each judgement himself. The examiner expects you to separate assistance (language, formatting, a checklist) from substitution (the assessment itself), and to ask whether paucity of time justifies cutting the judgement or only seeking more time. That earns its own short part.
Each point should name the value, show where the act strains it, and say what the value would have required. The last column is where the hidden demand is met.
| Value | Where the act strains it | What the value required |
|---|---|---|
| Accountability | The report carries his name, but its reasoning was drafted by a system he cannot fully trace. | Every judgement in the report checked against the thesis and defensible by him. |
| Integrity | The university receives an AI-generated report under his responsibility; whether he disclosed AI use or independently verified its substance is unspecified. | Honest authorship, disclosure where rules allow AI help, and the thesis kept confidential. |
1.3 The structure
The question names two values and carries one hidden demand. So the answer has three parts: accountability, integrity, and whether the excuse holds. The shares are a teaching guide, not a UPSC rule: the quality of reasoning decides the final allocation.
| Personal responsibility | The university appointed him, not a tool; the recommendation is his, whoever drafted it. |
| Verification | AI can misread methods, invent references or praise what is weak; he must check every judgement against the thesis. |
| Answerability | The candidate, the viva board and the university rely on reasons he can explain and defend. |
| Honesty and transparency | Presenting AI-generated assessment as wholly independent judgement misrepresents its origin; non-disclosure, where disclosure is required, compounds the breach. |
| Diligence and judgement | Years of a scholar’s work deserve a full reading; academic judgement is the very service he was asked for. |
| Confidentiality and fairness | If unpublished thesis material reaches an unapproved external tool, confidentiality may be compromised; if unverified AI conclusions are relied on, the scholar may receive an inaccurate or unfair assessment. |
Time pressure may justify requesting an extension, rescheduling, or using permitted assistance. It cannot justify submitting an assessment without adequate independent scrutiny and professional judgement.
1.4 How to present it
The form follows the question. “Discuss from the perspective of accountability and integrity” names two lenses, so the clearest form is points under two headings, each point naming the strain on that value. A brief third heading tests whether time pressure and partial editing justify the act; this can also be woven into the two parts, provided the structure stays clear.
| Form | Fit here | Reason |
|---|---|---|
| Points under two headings | Best | Keeps the two values apart, so the examiner sees both lenses applied in full. |
| Two-column table | Good | Accountability and integrity side by side; works if each cell stays short. |
| Flowchart | Weak | There is no process to show; the question asks for judgement, not sequence. |
1.5 The introduction and the conclusion
Keep this decode with you. Download the complete decode of this question as a free, print-ready PDF. A one-time sign-up opens every free download on the site.
Download the free PDFPYQ Intelligence 2013 to 2026
The question’s context in 2026, its place in a line of questions on accountability, integrity and technology, and what to prepare next.
2.1 Why the subject was topical in 2026
By 2026, research funders had issued specific restrictions on generative AI in peer review, while governments and international bodies had set broader AI governance principles. These are ethical parallels; they do not directly govern Ph.D. examination in Indian universities, where university rules apply. UPSC never states why it sets a question, so treat this as context, not cause.
| When | What happened | Why it matters for this question |
|---|---|---|
| November 2021 | UNESCO adopts its Recommendation on the Ethics of Artificial Intelligence, with human oversight and responsibility among its principles. | Responsibility for decisions must remain attributable to persons, not systems. |
| 23 June 2023 | The US National Institutes of Health prohibits its peer reviewers from using generative AI to analyse applications or draft critiques (NOT-OD-23-149). | The closest parallel: an evaluator’s judgement and the confidentiality of the work evaluated. |
| 14 December 2023 | The US National Science Foundation bars reviewers from uploading proposal content to non-approved AI tools; breach treated as a violation of its confidentiality pledge. | Confidentiality obligations extend to AI tools. |
| 5 November 2025 | MeitY releases the India AI Governance Guidelines under the IndiaAI Mission; accountability and “people first” are among its seven principles. | India’s own frame: AI use with human accountability. |
2.2 The skill UPSC keeps testing
These questions from 2013 to 2026 share one concern: what a value demands of a person in a concrete role, and whether technology changes that demand.
| Year | The question as UPSC set it, or a marked case summary | What it asked of you | What changed in 2026 |
|---|---|---|---|
| 2013 GS IV, Q1 | What do you understand by ‘values’ and ‘ethics’? In what way is it important to be ethical along with being professionally competent? | Separate competence from moral responsibility. | A competent professor can still fail ethically: competence is not the test. |
| 2014 GS IV, Q7 | What does ‘accountability’ mean in the context of public service? What measures can be adopted to ensure individual and collective accountability of public servants? | Define accountability and the mechanisms that secure it. | The same value, applied to an academic examiner using AI. |
| 2014 GS IV, Q3(a) | Question summary, not verbatim: interpret the statement, “Integrity without knowledge is weak and useless, but knowledge without integrity is dangerous and dreadful,” and explain its contemporary relevance with illustrations. | Relate knowledge and integrity. | The professor has the knowledge; the question tests his integrity. |
| 2015 GS IV, Q1(b)(iv) | Part summary, not verbatim: one of four pairs to differentiate, personal ethics and professional ethics. | The duties a role adds to personal morality. | An examiner’s duties are professional, owed to the scholar and the university. |
| 2017 GS IV, Q4(a) | One of the tests of integrity is complete refusal to be compromised. Explain with reference to a real life example. | Integrity under temptation. | Time pressure as the temptation to compromise. |
| 2021 GS IV, Q6(b) | ‘Integrity is a value that empowers the human being.’ Justify with suitable illustration. | Show integrity at work in a person’s conduct. | Integrity tested where a shortcut is easy and hard to detect. |
| 2021 GS IV, Q2(a) | Impact of digital technology as reliable source of input for rational decision making is a debatable issue. Critically evaluate with suitable example. | Weigh technology as an input to decisions. | From technology as an input to technology replacing the decision-maker’s judgement. |
| 2024 GS IV, Q1(a) | The application of Artificial Intelligence as a dependable source of input for administrative rational decision-making is a debatable issue. Critically examine the statement from the ethical point of view. | The ethics of relying on AI in administration. | The same concern in a university, with a named person and named values. |
| 2025 GS IV, Q1(a) | In the present digital age, social media has revolutionised our way of communication and interaction. However, it has raised several ethical issues and challenges. Describe the key ethical dilemmas in this regard. | Ethical dilemmas of digital technology. | Digital ethics tested again the year before, now inside a profession. |
| 2026 GS IV, Q8 | Case summary, not verbatim: an SP’s AI-enabled predictive policing system; asks the ethical issues including biases, the alternatives, and the most ethical action. | Accountability for AI-assisted decisions in policing. | The same paper tested technology and responsibility three times: Q1(a), Q8 and Q10(b). |
| 2026 GS IV, Q10(b) | A subpart of a 20-mark Section B case study: “Discuss the ethics of using technology as a gatekeeper for welfare schemes.” The case concerns biometric authentication failures, unreliable connectivity and exclusion from the Public Distribution System; the subpart must be read in that context. | Human responsibility for technology-mediated decisions. | The third 2026 question on technology and responsibility. |
- Define each value in one line.
- Apply each separately to the act.
- Name the duty each value creates for this role.
- Test the excuse the question offers.
- Give a verdict both values support.
2.3 What to prepare next
Values around this question
| Probity | Integrity shown in the handling of public trust and resources. |
| Transparency | Openness about how a decision was made, including the tools used. |
| Conflict of interest | Examiner duties when personal convenience competes with duty. |
| Professional ethics | Codes of conduct for doctors, teachers, examiners and researchers. |
AI and ethics, likely next
| Algorithmic accountability | Who answers when an automated decision harms someone. |
| Human in the loop | Meaningful human review, not a signature on a machine’s output. |
| Technology in welfare and policing | Q8 (2026): AI in predictive policing. Q10 (2026): biometric authentication in PDS and technology as a gatekeeper. |
| Academic integrity | AI use by students and scholars, and disclosure rules. |
Additional information
Everything needed to answer this question well. Tier 1 holds what the demand needs; Tier 2 places it within the wider debate on AI, responsibility and academic ethics.
| Subject | What it is | Why it matters for the answer |
|---|---|---|
| The two values | ||
| Accountability | The obligation to answer for one’s decisions to those affected and to those who entrusted the task, and to accept scrutiny of them. The Nolan Committee (UK, 1995) lists it among the Seven Principles of Public Life. | Part 1: the professor answers for the report, whoever drafted it. |
| Integrity | Consistency between one’s principles and one’s conduct, including when no one is watching; honesty in what one presents as one’s own. | Part 2: presenting a machine’s substantive assessment as independently formed would break that consistency. |
| How they differ | Accountability is answerability to others, enforced from outside; integrity is fidelity to one’s own principles, held from within. | Keeping the two apart is what the question tests. |
| The examiner’s role | ||
| Thesis evaluation in India | Regulation 11(5) of the UGC Ph.D. Regulations, 2022 (notified 7 November 2022) requires evaluation by the research supervisor and at least two external subject experts not employed by the institution; under Regulation 11(6), the viva voce is conducted when both external examiners recommend acceptance, after incorporating any corrections they suggest; if one recommends rejection, the thesis goes to an alternate external examiner from the approved panel, and the viva proceeds only if that examiner recommends acceptance. | The report is a formal step on which the award of a degree depends. |
| What an examiner owes | A full reading; judgement on originality, method and contribution; reasons for the recommendation; confidentiality of unpublished work; timeliness. | Each duty maps onto one of the two values. |
| Academic integrity rules | The UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 (notification dated 23 July 2018; published in the Gazette of India on 31 July 2018, the date from which they came into force) set standards and graded penalties for plagiarism in academic and research work. | Relevant context: they do not expressly cover AI-generated examiner reports, so the failure here is professional and ethical first; individual university rules may decide the rest. |
| Where AI enters | ||
| What AI can do | Improve language and formatting, suggest a report structure. Summarising confidential thesis material needs a tool and data arrangement the institution has authorised. | Legitimate assistance, where permitted and disclosed as required. |
| What AI cannot do for him | AI may help compare literature or flag methodological issues, but it cannot discharge the appointed examiner’s duty or bear accountability for the recommendation; it can also fabricate references and make confident errors, so every finding needs expert scrutiny. | The core of the accountability part. |
| Confidentiality | Uploading an unpublished thesis to an external tool may let the text be stored or used beyond the evaluation; in US grant peer review, NIH prohibits reviewers from using generative AI to analyse applications or formulate critiques, and NSF prohibits uploading proposal content and review records to non-approved AI tools. These rules govern their own grant review, not Ph.D. evaluation in Indian universities. | The confidentiality point in the integrity part. |
Where the line falls: assistance or substitution
Thinkers and ideas that fit
| Kant | Treat persons as ends, not merely as means: the scholar’s dignity and academic interests deserve a conscientious assessment, not one cut short for the examiner’s convenience. |
| Gandhi | Among the Seven Social Sins: knowledge without character. Expertise used without care fails its purpose. |
| Virtue ethics | Integrity as a settled character: what a good examiner does when the shortcut is easy and unseen. |
| Role morality | A role brings duties beyond ordinary ones: an examiner holds a gatekeeping trust for the university and the discipline. |
What AI use can legitimately offer
- Language: clearer reports, especially for non-native writers.
- Structure: a checklist so no criterion is missed.
- Speed: summaries to navigate a long thesis.
- Consistency: comparable reports across examiners.
What it puts at risk
- Judgement: conclusions not formed by the expert appointed.
- Accuracy: invented references, misread methods.
- Confidentiality: unpublished work sent to external systems.
- Trust: in the degree, the examiner and the university.
- Fairness: a candidate judged by a tool’s reading.
Two debates worth knowing
One view holds that any AI use corrupts expert judgement and risks confidentiality. The other holds that a ban is unenforceable and wastes useful aid. The balanced reading: permit assistance under clear rules and disclosure; never permit substitution of judgement.
Disclosure is an important part of transparency where AI assistance is permitted, but disclosure alone does not make unauthorised use or delegated academic judgement ethical. Integrity also requires diligence, confidentiality and compliance with applicable rules; accountability requires independent verification and a defensible recommendation. The balanced reading: where rules require disclosure, it is necessary but not sufficient; confidentiality, verification and independent judgement must follow.
Pitfalls and variations
The mistakes that cost marks here, the claims never to make, and how the answer changes when the question changes.
4.1 Six pitfalls, each with its correction
4.2 Never write these
- “UGC has banned AI in thesis evaluation.” No such specific rule is established; argue from the values, the examiner’s duties and university rules.
- “The professor uploaded the whole thesis and hid the AI use.” The question does not say so; reason conditionally.
- “Using AI is always unethical.” Assistance can be legitimate when authorised, confidential, disclosed where required and independently verified.
- “AI evaluation is more objective than a human examiner.” AI can err and fabricate; it bears no responsibility.
- “The professor should be dismissed.” The question asks for an ethical discussion, not a penalty.
- “AI tools keep uploaded documents private.” Many give no such guarantee; that is the confidentiality risk.
4.3 If the question changes
| If it asks | The structure becomes | What must change |
|---|---|---|
| What are the ethical issues involved? | A list of issues: honesty, responsibility, confidentiality, fairness, professional duty | No fixed lenses; group the issues and rank them. |
| What should the university do? | 1. Immediate steps on this report; 2. A policy on AI in evaluation | Fresh evaluation, disclosure rules, permitted and barred uses. |
| The same act by a civil servant drafting a file note | The same two lenses, with public accountability and official secrecy | Rules of business and data protection replace university rules. |
| As a 20-mark case study | Stakeholders; ethical issues; options with merits and demerits; the course chosen | Options: withdraw the report, disclose, seek extension, re-evaluate. |
The model answer
One answer, written in 150 words, in the form this question calls for: points under the two named values, with a short part for the hidden demand. The introduction covers the whole question; the conclusion settles its demand.
Accountability requires an examiner to justify decisions; integrity requires fidelity to professional duties. A doctoral evaluation affects a scholar’s degree.
1. Accountability
- Personal responsibility: the university appointed the professor, not an AI system; the recommendation remains his.
- Verification: AI may invent citations or misunderstand methods. Editing a generated report cannot establish that he independently examined the thesis before making a reasoned recommendation.
- Answerability: he must substantiate every conclusion to the university and candidate.
2. Integrity
- Honesty: concealing substantive AI assistance, especially where disclosure is required, compromises transparency.
- Diligence: independent scrutiny of originality, methodology and contribution cannot be outsourced.
- Fairness and confidentiality: unverified criticism may prejudice the scholar; sharing unpublished material with unauthorised tools risks exposure.
3. Can time justify it?
Time pressure warrants requesting an extension or using approved language assistance, not delegating academic judgement.
Thus, AI may support drafting, but defensible evaluation, ethical disclosure and responsibility for the judgement must remain human.
Other ways to write it
The model answer above remains the answer. These outlines show how a different, equally valid answer could be built, and when each choice works.
| Approach | How it is built | When it works |
|---|---|---|
| Two-column table | Rows for responsibility, honesty, diligence, confidentiality; columns for accountability and integrity. | When you want the contrast between the values to show at a glance. |
| Open with the stakeholders | Introduction names the candidate, university and discipline; then the two values. | When fairness to the candidate is the point you want to lead with. |
| Assistance against substitution | Part 3 built around the test: can the examiner independently substantiate every substantive finding through direct examination of the thesis, without treating AI-generated conclusions as authoritative? | When the hidden demand is your strongest point. |
| As a case study | Options for the professor now: withdraw and re-evaluate, disclose, seek time; the course chosen. | When the question asks what he should do. |
Quick recall and Prelims crossover
Facts for rapid revision and Prelims.
| 1 | Accountability: answerability for decisions to those affected and those who entrusted them. | 9 | NIH barred peer reviewers from using generative AI: 23 June 2023. |
| 2 | Integrity: consistency between principles and conduct. | 10 | NSF notice on AI in merit review: 14 December 2023. |
| 3 | Nolan Principles (UK, 1995): seven principles of public life. | 11 | UNESCO Recommendation on the Ethics of AI: November 2021. |
| 4 | Nolan list: selflessness, integrity, objectivity, accountability, openness, honesty, leadership. | 12 | India AI Governance Guidelines (MeitY): 5 November 2025. |
| 5 | UGC Ph.D. Regulations, 2022: supervisor plus at least two external examiners. | 13 | Its seven principles include accountability and people first. |
| 6 | UGC academic integrity regulations: 2018, graded penalties for plagiarism. | 14 | EU AI Act, Article 6 and Annex III, point 3(b): certain AI systems intended to evaluate learning outcomes are high-risk, subject to its conditions and exceptions; this EU rule does not automatically govern Ph.D. evaluation in India; its application depends on the Act’s territorial scope and the AI system’s intended use. |
| 7 | Second ARC, Fourth Report: Ethics in Governance (2007). | 15 | Gandhi’s Seven Social Sins include knowledge without character. |
| 8 | GS IV PYQs on technology: 2024 (AI in administration), 2025 Q1(a) (social media), 2026 Q1(a), Q8, Q10(b). | 16 | Kant: treat persons as ends, never merely as means. |
Write it yourself
Write the answer in 150 words without looking back, then check it against each line below.
| Check | Yes / No | |
|---|---|---|
| 1 | Does the introduction define both values and say why this act engages them? | |
| 2 | Are accountability and integrity under separate headings, each applied to the act? | |
| 3 | Does Part 1 show that editing the report is not evaluating the thesis? | |
| 4 | Does Part 2 cover honesty, diligence and confidentiality? | |
| 5 | Does Part 3 test the excuse of paucity of time and separate assistance from substitution? | |
| 6 | Does the conclusion give a verdict both values support, at about the length of the introduction? | |
| 7 | Is the answer within the prescribed 150-word limit, headings included, and free from unsupported assumptions about the professor’s conduct (section 4.2)? |
Write your first answer. Get written feedback, free.
Reading a decode helps you understand the question. Writing your own answer helps you put that understanding into practice. Write this answer and send it to us: your first answer receives a free written evaluation by the UPSC Gurus method.
- Write the answer by hand on A4 sheets, within about 150 words, before you look at the model answer.
- Go through the self-check in section 8 above, then photograph every page clearly (or scan it as one PDF).
- Upload it below. Your written feedback comes to your email: how your answer met the question’s demand and implied dimensions, its structure, content and presentation, and what to work on next.
Explore personalised support
Write any actual UPSC Mains GS paper from 2017 to 2026. Receive written evaluation of every answer by email, followed by a personal feedback session.
Get my paper evaluatedUnderstand where your preparation stands through a structured diagnosis, a written roadmap and one follow-up session. The full counselling fee is adjusted against any UPSC Gurus programme you join within 30 days. This is a fee adjustment, not a refund.
Book counsellingSources
- University Grants Commission: UGC (Minimum Standards and Procedures for Award of Ph.D. Degree) Regulations, 2022, Regulation 11(5)-(6), notified 7 November 2022; UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018, notification dated 23 July 2018, published in the Gazette of India on 31 July 2018.
- European Union, Regulation (EU) 2024/1689 (AI Act), Article 6 and Annex III.
- National Institutes of Health, Notice NOT-OD-23-149 (23 June 2023); National Science Foundation, Notice to the research community on generative AI in merit review (14 December 2023).
- UNESCO, Recommendation on the Ethics of Artificial Intelligence (November 2021).
- Ministry of Electronics and Information Technology, India AI Governance Guidelines (5 November 2025).
- Committee on Standards in Public Life (UK), Seven Principles of Public Life (1995); Second Administrative Reforms Commission, Fourth Report, Ethics in Governance (2007).
- UPSC Civil Services (Main), General Studies Paper IV, official question papers and archives (upsc.gov.in): 2013 Q1; 2014 Q3(a), Q7; 2015 Q1(b)(iv); 2017 Q4(a); 2021 Q2(a), Q6(b); 2024 Q1(a); 2025 Q1(a); 2026 Q1(a), Q8, Q10(b).
Your first Mentor’s Desk call is free
One free conversation per aspirant about where you are in your preparation, with an honest suggestion of the right next step.
Know more and bookHave a question? Get in touch
For questions about a programme, its fees, batch dates, payment or access. Message us on WhatsApp and our team will reply. Asking a question is not a booking, and you can ask as often as you need.
Get in touch on WhatsApp
