Why a prompting workshop is not an AI-literacy strategy
Knowing how to ask a generative AI system for an answer is useful. It does not establish whether that answer is true, ethically obtained, appropriate for a learner, or safe to use in an academic decision. A university can train hundreds of people to write prompts without establishing who is responsible when an AI-generated citation is invented, a student's private record is exposed, or automated feedback rewards fluent but incorrect work.
AI literacy is a combination of operational skill, informed judgement and institutional safeguards. For Pakistani universities, those capabilities must be adaptable to different language preferences, connectivity, disciplinary cultures and access to paid tools. The goal is not the largest number of tools adopted; it is trustworthy, inclusive educational practice.
Five capabilities every university should teach
1. Frame the task before prompting
Define the educational purpose, intended audience, constraints and what an acceptable outcome would look like. Learners should compare an AI-assisted route with a non-AI alternative rather than assume the tool is always needed.
2. Verify claims, citations and calculations
Students and faculty should trace consequential statements to primary sources, check whether quoted references exist, and test numbers independently. A polished response is not a substitute for evidence. Verification needs to be practised in discipline-specific assignments rather than presented as a warning slide.
3. Protect information and academic integrity
Teach clear boundaries for unpublished research, assessment material, personal records and confidential organisational data. Each course should say what AI assistance is allowed, how it must be disclosed, and which work must demonstrate the learner's own reasoning.
4. Design learning rather than outsource thinking
Faculty can use AI for formative examples, alternative explanations and feedback drafts. Assessment should still observe the learner's actual competence through explanation, source critique, reflection, oral defence or applied work where appropriate. Accessibility and non-AI alternatives matter when access is unequal.
5. Keep accountable humans in charge
Institutions need named reviewers for high-impact uses such as admissions, grading or student support. The person reviewing an output must have the time, authority and knowledge to reject it. Document how students and employees can question consequential decisions.
A 30-day pilot for a department
- Week 1: Map the actual use. Interview students and faculty about tasks already using AI. Identify sensitive information and where decisions affect people.
- Week 2: Run a small skills clinic. Use one authentic task to practise prompting, verification, disclosure and non-AI alternatives. Document the baseline.
- Week 3: Redesign one assessment or workflow. State permitted use, required evidence and the human review point. Test with a small, diverse learner group.
- Week 4: Evaluate and decide. Compare quality, errors, workload, accessibility and student understanding against the baseline; decide to revise, extend or stop the pilot.
Measure whether people can explain and defend their AI-assisted work, not simply whether they can produce it faster.
From individual skills to institutional readiness
Training only end users leaves gaps in procurement, policy, data governance and leadership oversight. Fiker Futures Academy's PAUSE-AI checklist provides six questions for evaluating an AI initiative: Purpose, Accountability, Users and equity, Safety, Evidence, and Alternative futures. It is a practical decision aid, not a certification standard.
Use the PAUSE-AI Responsible AI Readiness Checklist before scaling a pilot. For a plain-language overview of the topic, read AI literacy in Pakistan.
What an effective learning outcome looks like
- A student can identify an unsupported assertion and find a reliable source to confirm or correct it.
- An educator can distinguish an acceptable AI-assisted draft from work that fails the course's assessment or disclosure requirements.
- A department can name the owner, safeguards, evaluation measures and stop conditions for its AI pilot.
- An institution can provide accessible alternatives for learners who cannot use a particular paid tool.
Explore the training
Dr. Salman Ahmed Khatani, Associate Professor and founder of Fiker Futures Academy in Karachi, works with educators, students and organisations on AI literacy and responsible, human-centred adoption. Explore the AI course instructor profile, review available programmes or request a university workshop. Programme dates, instructors and formats should be confirmed with the academy before enrolment.