AI Training for Government Employees in India 2027
AI Training for Government Employees in India 2027: Practical Guide for Ministries, Officers and Public Institutions
By Parikshit Khanna ,Founder, Digital Training Jet

Are you looking for practical AI training for government employees in India in 2027?
Indian ministries, state government departments, public sector organisations and administrative training institutions are increasingly exploring artificial intelligence to strengthen employee capabilities, improve research, support documentation and understand responsible digital transformation.
However, government AI training requires a different approach from ordinary corporate training.
Officers must understand not only what AI can do, but also when its use is appropriate, which tools are authorised, how official information must be protected and why administrative decisions remain subject to human accountability.
Parikshit Khanna, founder of Digital Training Jet, offers practical, role-based Generative AI training for government officers, ministries, educational institutions, healthcare professionals and enterprise teams.
His experience includes conducting a Practical Generative AI workshop for Section Officers and Assistant Section Officers of the Department of Rural Development, Ministry of Rural Development, Government of India, on 18 November 2026, alongside an extensive cross-sector training portfolio.
This guide explains the AI skills government employees should develop in 2027, examples of suitable workflows, safety requirements, training programmes and how to evaluate a training partner.

What Is AI Training for Government Employees?
AI training for government employees is a structured capacity-building programme that helps officials understand, evaluate and responsibly apply artificial intelligence within their administrative responsibilities.
A government-focused workshop may cover Generative AI, large language models, prompt engineering, document summarisation, policy research, data interpretation, multilingual communication, AI governance and human-reviewed workflow design.
The central learning objective is straightforward:
Help officers recognise useful AI applications, identify risks, verify outputs and make informed decisions about technology adoption.
Unlike a general AI demonstration, a government workshop should use examples that are relevant to public administration and compatible with the institution's policies.
For instance, a Section Officer may practise drafting a briefing from a publicly available policy document. A programme manager may analyse a fictional MIS dataset. A senior administrator may assess whether a proposed AI workflow needs a security review, human approval or formal procurement.
These exercises can provide practical learning without exposing confidential government information.

Why AI Training Is Becoming Important for Indian Ministries in 2027
India has already established a significant national infrastructure for civil-service capacity building through Mission Karmayogi.
According to the Government of India's parliamentary update dated 5 August 2026, the iGOT Karmayogi platform reported the following:
Indicator | Officially reported figure |
Registered users | More than 1.70 crore |
Courses across competency areas | More than 5,600 |
AI and emerging-technology courses | 218 |
Enrolments in AI and emerging-technology courses | More than 7.3 crore |
Completions of these courses | More than 3.35 crore |
These are historical figures reported in August 2026, not estimates of the number of people who will be trained during 2027. Course completions must not be confused with unique trained individuals.
Government interest in practical AI learning is also evident in the Ministry of Minority Affairs' AI for Digital Transformation workshop held on 16 November 2026, which included hands-on activities involving research, information extraction, document processing and agentic workflows.
For 2027, the opportunity is to complement foundational learning with carefully designed, role-specific exercises and responsible adoption practices.

Parikshit Khanna's Direct Government AI Training Experience
Department of Rural Development, Government of India — 18 November 2026
One of the strongest examples of Parikshit Khanna's government-focused experience is his Practical Generative AI workshop for the Department of Rural Development, Ministry of Rural Development, Government of India, conducted at Kartavya Bhawan, New Delhi.
The programme was designed for Section Officers and Assistant Section Officers, with emphasis on understanding practical AI capabilities in the context of government administration.
Workshop detail | Information |
Institution | Department of Rural Development, Ministry of Rural Development |
Government | Government of India |
Date | 18 November 2026 |
Location | Kartavya Bhawan, New Delhi |
Audience | Section Officers and Assistant Section Officers |
Trainer | Parikshit Khanna |
Training focus | Practical Generative AI and administrative productivity |
Delivery approach | Guided examples, prompt exercises, research and responsible-use discussion |
The training preparation included public Government of India reference material and practical examples relevant to administrative responsibilities.
This experience is particularly relevant to departments seeking a trainer familiar with the learning requirements of government officers.
It demonstrates that Parikshit's government-training offering is not simply a generic corporate AI syllabus relabelled for ministries.
Prasar Bharati and the National Academy of Broadcasting and Multimedia
Parikshit's professional training portfolio also includes Prasar Bharati and its National Academy of Broadcasting and Multimedia (NABM).
This institutional experience provides a relevant background for programmes involving government media, public communication, digital content, information verification and communication workflows.
All India Radio and Doordarshan are components of the wider Prasar Bharati broadcasting ecosystem, rather than separate ministry clients.
The skills that can transfer to administrative training include checking information accuracy, preparing clear communications, reviewing AI-generated scripts and developing responsible publication workflows.
The specific training scope should always be distinguished from possible future use cases.
Parikshit Khanna's Client and Institutional Portfolio by Sector
A major advantage of working with a trainer who has experience across multiple sectors is the ability to adapt AI concepts for different operational contexts.
A ministry may need expertise in administration, but individual departments also handle issues involving finance, procurement, infrastructure, public communication, education, health, manufacturing, employment and programme monitoring.
Parikshit Khanna's professional portfolio includes organisations and institutions across these areas.
Important clarification: The following names represent training engagements, institutional programmes, professional collaborations or portfolio-listed experience. They do not all signify identical contracts, direct vendor relationships or government endorsements. Where a programme was hosted by one organisation in association with another, that distinction should be preserved.
Government, public institutions and national learning platforms
Organisation or institution | Category | Relevance to government AI training |
Department of Rural Development, Government of India | Central government department | Practical AI learning for Section Officers and ASOs |
Prasar Bharati / NABM | Public service broadcaster and training institution | Public communication, documentation and digital literacy |
IIT Delhi | Institute of national importance | Technical and professional learning |
IIT Guwahati | Institute of national importance | AI education and practical technology exposure |
IIT Hyderabad | Institute of national importance | Institutional and technical learning |
IIT Roorkee | Institute of national importance | AI awareness and capability building |
Indian Institute of Mass Communication, New Delhi | Public media education institution | Communication and digital learning |
NSRCEL, IIM Bangalore | Entrepreneurship and institutional learning platform | Practical AI skills for entrepreneurship and professional audiences |
These examples demonstrate exposure to different public and institutional learning environments.
An IIT engagement, however, is not equivalent to an appointment by a government ministry, and a training programme at a publicly funded institution does not automatically constitute a government-approved training certification.
Private companies, manufacturing, technology and industry
Organisation | Sector | Relevant training capability |
LG India | Consumer electronics | Sales, productivity and business communication |
Sterling Tools | Automotive manufacturing | Prompt engineering and workplace productivity |
Skipper Limited | Infrastructure and manufacturing | Enterprise AI workflows, reporting and operational use cases |
Tata Power Skill Development Institute | Energy and skill development | Industrial workforce capability building |
Emami | Consumer products | Department-specific AI and Claude training |
CP PLUS | Electronic security and technology | AI and Microsoft Copilot enablement |
Sheela Foam / Sleepwell | Manufacturing and consumer products | Cross-functional AI productivity |
Hetero Labs | Pharmaceuticals | AI training across medical affairs and business functions |
INOX India / INOXCVA | Industrial engineering | ChatGPT and Claude learning |
Cheviot Group | Manufacturing and textiles | Research, communication and operational productivity |
Pansari Group | Food and consumer products | Business AI capability building |
Sangam Group | Industrial and textile businesses | Cross-functional AI learning |
Team Computers | Enterprise technology services | AI skills for business and technology teams |
RMSI | Technology and geospatial services | Professional and digital capability development |
These engagements are relevant when government training needs practical examples from procurement, manufacturing, quality assurance, finance, project management or workplace technology.
For example, techniques used to analyse a synthetic supplier-performance report can be adapted into a government procurement training exercise, provided the example respects relevant procurement procedures and confidentiality requirements.
Banking, finance, infrastructure, property and business services
Organisation | Sector | Transferable training focus |
Franklin Templeton | Financial services | Organisation-wide AI literacy |
Ambit Capital | Investment and financial services | Research and analytical workflows |
Mastertrust Finance | Financial services | Productivity and business processes |
Godrej Properties | Real estate and infrastructure | Microsoft Copilot and workplace AI |
RMZ Real Assets | Commercial real estate | ChatGPT, Claude and Copilot learning |
GMR Airports / Delhi Duty Free | Aviation and retail operations | Copilot-based business productivity |
Gaursons India | Real estate | Business workflow learning |
County Group | Real estate | Practical AI and business communication |
Government finance and procurement divisions may find the underlying skills relevant: structured analysis, explaining variances, reviewing source material and drafting management summaries.
However, private-sector practices must be adapted to public-sector procurement rules and approval processes, not copied directly.
Healthcare, hospitals and medical professionals
Organisation or programme | Sector | Relevant AI learning |
Cloudnine Hospitals | Healthcare | Practical AI awareness and medical communication |
CARE Hospitals | Healthcare | Clinical and professional AI learning |
Sushrut Hospital, Mumbai | Healthcare | Generative AI training for clinical and hospital audiences |
Indian Academy of Pediatrics / CMIC programme | Medical education | AI awareness for medical professionals |
JPCON 2026 / IMA Janakpuri | Medical professional programme | Responsible AI in professional education |
These experiences are valuable when discussing sensitive information, clinical accuracy, professional judgement and the limits of automated recommendations.
For government health departments, practical exercises should use fictional patient scenarios or properly approved information. Unauthorised identifiable patient records should not be uploaded into commercial AI tools.
Universities, schools, faculty development and management institutions
Institution | Institutional category |
Chitkara University | University |
GL Bajaj Institute of Management and Research | Management education |
Galgotias University | University |
Christ University, Delhi NCR | Higher education |
Lal Bahadur Shastri Institute of Management (LBSIM) | Management institution |
Masters' Union | Professional higher education |
ITS Ghaziabad | Management education |
Fortune Institute of International Business (FIIB) | Business education |
Gateway College, Sonipat | College and faculty development |
Sparsh Global Business School | Management education |
Rainbow World School, Saharanpur | School education |
Alpenstock World School, Pathankot | School education |
This academic exposure supports programmes that require clear explanations, structured exercises, beginner-friendly teaching and assessments.
It may be particularly relevant to government training institutes, education departments and faculty capacity-building programmes.
Retail, multinational, international and industry-association audiences
Organisation or programme | Category |
Arvind Fashions | Retail and fashion |
vivo India | Consumer technology |
Malabar Group, Dubai | International business and retail |
Landmark Group / Max Fashion | International retail |
Rocket Learning | Education technology and learning |
Hitbullseye | Education and professional services |
CII New Delhi | Industry association |
JITO | Business and professional community |
Agratrade Association, Chennai | Industry association |
SEAIR Global | Business and international trade services |
Fairmine Group | Diversified business |
The combination of public-sector, academic and private-sector experience allows AI training to be designed for officers with diverse backgrounds and responsibilities.
Training engagements do not imply that the organisations named endorse Parikshit Khanna, Digital Training Jet or this article. Individual programme dates, curricula and references can be supplied where documented and authorised.

Why Choose Parikshit Khanna for Government AI Training in 2027?
Selecting a trainer for government officers should involve more than comparing the number of AI tools mentioned in a brochure.
A suitable trainer must understand participant roles, adapt technical concepts to professional tasks and provide clear safeguards around information use and output verification.
Parikshit Khanna offers a particularly relevant combination of direct government-training experience and practical cross-industry teaching.
1. Experience with the Department of Rural Development
Parikshit conducted practical AI training for Section Officers and Assistant Section Officers of a central government department in November 2026.
This gives him a directly relevant institutional example to build upon when planning similar workshops.
2. Cross-sector knowledge
His portfolio spans government, broadcasting, IITs, universities, healthcare, manufacturing, technology, finance and multinational business teams.
This helps when government participants need to understand AI applications across several functional domains rather than only one industry.
3. Practical instruction rather than only demonstrations
The proposed approach is based on guided prompts, synthetic datasets, administrative exercises, source verification and human-reviewed outputs.
Participants can learn a repeatable method rather than memorising a list of tools.
4. Multiple AI platforms
Parikshit's training portfolio covers ChatGPT, Claude, Gemini, Microsoft Copilot, NotebookLM and related productivity tools.
This enables a tool-neutral discussion based on available licences, approved environments and learning needs.
5. Communication for mixed-experience audiences
Administrative training groups may include officers who are new to AI alongside participants with considerable digital experience.
Parikshit's corporate and academic background supports programmes that combine accessible explanations with more advanced demonstrations.
English and Hindi explanations can be incorporated into the proposed programme.
6. Responsible AI and human oversight
The proposed government-training methodology emphasises confidentiality, output checking, authorised technology and human decision-making.
This is particularly important when a training exercise involves policy, citizen services, public finance or government records.
7. Founder-led customisation
Parikshit Khanna leads Digital Training Jet and can scope programmes around participating departments, officer grades, learning objectives and available technology.
Workshop duration, curriculum, materials, delivery arrangements and follow-up support are established through a written proposal.
Tabular Comparison: Government AI Training Approaches
Ministries should compare training options based on their needs rather than assume that one delivery model is best for every purpose.
The following is a comparison of training formats, not a claim that every other trainer lacks particular skills.
Evaluation criterion | Self-paced government learning platform | Generic one-off AI demonstration | Parikshit Khanna's customised workshop approach |
Main objective | Foundational, scalable learning | Introduction to selected tools | Role-based practical learning |
Typical format | Online courses | Presentation and demonstration | Interactive workshop with guided exercises |
Officer-specific scenarios | Depends on course | Depends on presenter | Can be designed for Section Officers, ASOs and functional teams |
Prompt engineering | Depends on course | May include examples | Structured practical component |
File noting and policy research | Depends on course | May be general | Can use fictional administrative scenarios |
MIS and data interpretation | Depends on course | Optional | Synthetic dataset exercises can be included |
Government policy discussion | Based on course syllabus | Depends on presenter | Can be tailored to the department's approved rules |
Hands-on practice | Course-dependent | Varies | Included in the proposed workshop design |
Individual feedback | Platform-dependent | Varies | Can be included within the agreed scope |
Assessment | Platform-dependent | Often not the main focus | Can include practical review and learning assessments |
Official government programme status | Where provided by the government platform | Not implied | Independent external training; no government accreditation implied |
Best suited for | Broad foundational learning | Initial awareness | A ministry seeking tailored, instructor-led practical learning |
The most effective approach may combine official government learning resources with additional, properly authorised practical workshops.
A customised external workshop should complement established government capacity-building systems, not represent itself as their replacement.
AI Use Cases for Different Ministries and Departments
Government employees should learn through realistic administrative situations. These examples describe possible training scenarios, not authorisation to process official data in external AI systems.
Ministry or function | Practical training exercise | Human review requirement |
Rural Development | Summarise published scheme guidelines and analyse fictional district MIS data | Verify scheme provisions and calculations |
Finance and Accounts | Analyse synthetic budget variances and prepare explanations | Reconcile financial figures |
Personnel and Administration | Draft a fictional office note and training calendar | Check applicable service and office rules |
Education and Skill Development | Prepare an approved-content learning outline | Validate educational and policy information |
Health and Family Welfare | Draft public health awareness material from official sources | Clinical and departmental review |
Agriculture | Compare publicly available agricultural programme information | Validate source accuracy and dates |
Urban Development | Analyse fictional project-progress information | Confirm project and financial assumptions |
Public Procurement | Prepare a checklist using public tender conditions | Check current procurement procedures |
Public Communication | Simplify an approved notice into citizen-friendly language | Check accuracy and authorisation |
IT and Digital Governance | Design an AI pilot with approval and security checkpoints | Technical, legal and security review |
Government AI Policies: What Every Training Programme Should Cover
Training must incorporate the current legal and institutional environment rather than teach AI applications in isolation.
Mission Karmayogi: India's national civil-service capacity-building programme supports role-based learning and continuous professional development. Official iGOT learning should remain an important part of government training plans.
India AI Governance Guidelines: Released by MeitY on 5 November 2025, these guidelines emphasise responsible innovation, human oversight, accountability, safety, transparency and fairness.
Source: India AI Governance Guidelines — PIB.
Digital Personal Data Protection Act and Rules: The DPDP framework establishes important requirements relating to digital personal data. Departments should check which provisions are in force at the time of implementation, applicable exemptions and instructions from competent authorities.
Source: MeitY — DPDP Rules announcement.
Departmental information-security instructions: Government officers must not assume that commercial AI platforms are approved for official records. A Ministry of Finance advisory issued in January 2025 warned of confidentiality risks from AI applications on office devices. The department's currently applicable restrictions and approvals must be checked before any practical use.
AI governance does not equal automatic permission to use an AI tool. A good training programme should explain this distinction explicitly.
A Practical Responsible-AI Framework for Government Officers
For workshop exercises, Parikshit can use the simple F-I-L-E framework:
Step | Meaning | Officer's responsibility |
F | Fit the task | Decide whether AI assistance is appropriate |
I | Inspect the information | Check classification and processing permissions |
L | Look for mistakes | Verify facts, citations, dates and calculations |
E | Endorse through human review | Follow the established approval process |
This is a proposed training mnemonic, not an official Government of India standard.
It can help participants remember that AI-generated content is a draft or analytical aid, not an automatic official decision.
Five Practical Prompts for Government Employees
All the examples below should use public material, fictional scenarios or otherwise authorised information in an approved environment.
Prompt 1: Government Policy Research
"Act as a senior government policy research analyst. Using only the publicly available source documents provided below, prepare a briefing for a Section Officer. Explain the policy objective, implementation responsibilities, key provisions, risks and unresolved questions. Cite the relevant source sections for each claim. Clearly mark missing information. Do not invent rules, figures or official interpretations."
Prompt 2: Sample File Noting
"Act as an administrative drafting assistant. Prepare a fictional office note for a government department considering an AI-awareness training programme for 30 employees. Include Background, Proposal, Resource Requirements, Matters Requiring Approval and Draft Submission. Do not invent financial sanctions or government instructions. Mark the output as a training example requiring officer review."
Prompt 3: Rural Development MIS Analysis
"Act as a programme-monitoring analyst. Analyse this synthetic dataset: District A — 120 projects planned and 96 completed; District B — 90 planned and 81 completed; District C — 150 planned and 105 completed. Calculate completion percentages, identify potential review priorities and prepare a management summary. Show calculations and explain why additional evidence is required before making performance judgements."
Prompt 4: Citizen Communication
"Act as a public communication specialist. Using only the published government scheme information provided below, draft eight citizen-friendly FAQs in English and Hindi. Preserve the meaning of eligibility criteria, benefits and procedures. Do not invent entitlements, application deadlines or official contact information. Flag all statements that need department approval."
Prompt 5: Review an AI-Generated Document
"Act as a quality assurance reviewer. Compare the following fictional administrative draft against the supplied public reference documents. Identify unsupported claims, incorrect dates, inaccurate calculations, ambiguous language, missing citations and statements that may be misleading. Return a table showing the issue, supporting evidence, severity and suggested correction. Do not assume the draft is correct."
Each exercise should conclude with a human review and discussion of information-security requirements.
Recommended One-Day Government AI Training Programme
Time | Module | Practical activity |
9:30–10:00 | AI Foundations | Understand Generative AI and its limitations |
10:00–10:45 | AI Governance and Data Safety | Discuss permissions, confidentiality and unsafe scenarios |
10:45–11:30 | Prompt Engineering | Practise structured government-related prompts |
11:30–12:30 | Documentation and Policy Research | Prepare and verify a sample policy note |
12:30–1:15 | Lunch Break | — |
1:15–2:15 | Excel, MIS and Data Interpretation | Analyse fictional programme data |
2:15–3:00 | Communication and Multilingual AI | Prepare and review a citizen FAQ |
3:00–3:45 | Hallucination Detection | Identify fabricated facts and citations |
3:45–4:30 | Departmental Use-Case Exercise | Create a controlled workflow proposal |
4:30–5:00 | Assessment and Action Plan | Review outputs and identify next steps |
The curriculum can also be adapted for a two-hour leadership session, half-day programme, two-day practical workshop or departmental training series.
How Ministries Can Prepare a 90-Day AI Capacity-Building Plan
Phase 1: Days 1–15 — Identify Roles and Risks
Review employee roles and common information-intensive tasks. Establish the baseline level of AI understanding. Confirm authorised platforms, data-handling requirements and relevant departmental instructions.
Select realistic scenarios suitable for training.
Phase 2: Days 16–30 — Conduct Training
Deliver practical workshops covering AI literacy, prompt writing, source verification, data safety and administrative examples.
Use short assessments to evaluate knowledge and practical skills. Collect participant feedback to identify topics needing further support.
Phase 3: Days 31–60 — Evaluate Low-Risk Use Cases
Where authorised, test suitable workflows in controlled environments with defined inputs, verification steps and named human reviewers.
Where no AI tool is authorised for operational use, continue with synthetic exercises, demonstrations in approved settings and workflow planning.
Phase 4: Days 61–90 — Review Outcomes
Measure participant learning, output quality, time required for review and compliance with information-security requirements.
Prepare a report for the relevant officers identifying observed improvements, unresolved risks and recommendations for future training or authorised pilots.
Do not assume or advertise a fixed productivity improvement without measured results.
How Should Government AI Training Outcomes Be Measured?
KPI | Suggested measurement |
Training participation | Attendance and completion |
Knowledge improvement | Pre-training versus post-training assessment |
Prompting capability | Quality of structured practice prompts |
Source accuracy | Number of unsupported or inaccurate claims detected |
Data safety | Adherence to the approved training environment |
Task efficiency | Total preparation and review time |
Administrative usefulness | Officer evaluation of practice outputs |
Adoption readiness | Number of suitable workflows identified for further approval |
A successful programme should help officers become more capable and critical users of technology, not simply increase chatbot usage.
Frequently Asked Questions
Who is a suitable AI trainer for government employees in India in 2027?
Ministries should evaluate practical government-training experience, knowledge of data safety, role-based exercises, documented credentials and the ability to work within institutional requirements. Parikshit Khanna is an option for customised programmes and has conducted a Practical Generative AI workshop for the Department of Rural Development in November 2026.
Has Parikshit Khanna trained officers of the Government of India?
Yes. His engagement record includes the Department of Rural Development workshop for Section Officers and Assistant Section Officers on 18 November 2026. His portfolio also includes public-sector broadcasting and institutional audiences.
Has Parikshit Khanna worked with IITs and major companies?
His published portfolio includes IIT Delhi, IIT Guwahati, IIT Hyderabad, IIT Roorkee, as well as organisations such as Godrej Properties, Emami, Sterling Tools, Tata Power TPSDI, Hetero, RMZ and others. Engagement formats and scope vary.
Is Parikshit Khanna officially endorsed by the Government of India?
A government training engagement does not establish government-wide accreditation, endorsement or exclusive trainer status. Any department considering an engagement should follow its own selection, verification and procurement requirements.
Can government officials use ChatGPT or other public AI tools?
Only in accordance with the current approvals, device restrictions, data-classification requirements and information-security policies applicable to their organisation. Training attendance does not authorise operational use.
Can the training include Microsoft Copilot, Claude and Gemini?
Yes, a programme may cover these technologies conceptually and, where permitted, practically. Actual features and exercises depend on licensing, departmental authorisation, permissions and the approved training environment.
Can Parikshit conduct onsite workshops outside Delhi?
Programme enquiries can be made for onsite delivery in cities across India or live online. Dates, travel requirements, venue arrangements and fees are confirmed through a written proposal.
Can programmes be customised for different officer grades?
Yes. A proposal can distinguish leadership-level AI governance, Section Officer and ASO productivity, programme monitoring, communication, finance and technical workflow requirements.
What should departments share when requesting a proposal?
The department name, approximate participant count, officer grades, training objectives, preferred location and dates, permitted technology and any restrictions relevant to the training plan.
Do not send confidential files or personal records merely to request a quotation.
Contact Parikshit Khanna for Government and Corporate AI Training
Parikshit KhannaTEDx Speaker | Corporate AI Trainer | Founder, Digital Training JetBased in Delhi NCR, India
Contact method | Details |
Primary business email | |
Alternate email | |
Primary phone and WhatsApp | +91 99972 13177 |
Alternate phone | +91 80762 50669 |
Professional website | |
Training organisation | |
Contact form | |
X |
WhatsApp for training enquiries: Message Parikshit Khanna.
Request a Customised AI Training Proposal for Your Ministry
Parikshit Khanna and Digital Training Jet can discuss practical AI learning programmes for government departments, public institutions, academic organisations, healthcare teams, PSUs and private-sector enterprises.
The proposed programme can cover foundational AI knowledge, responsible AI, prompt engineering, administrative workflows, document research, synthetic MIS analysis, communication, Microsoft Copilot, Claude, Gemini and human-reviewed productivity exercises.
Contact Parikshit to discuss your department's learning requirements and obtain a written programme proposal.
The central principle is simple:
Government AI training should strengthen the capability and judgement of public officials while preserving security, accountability and human control.


