Which ethical concept is often debated regarding the automated decision-making logic of autonomous vehicles?
- Logic Ambiguity
- Speed Threshold
- Trolley Problem
- Data Complexity
Explanation: The Trolley Problem thought experiment is used to analyze how AI should weigh different lives during unavoidable traffic accidents.
Which specific technical framework ensures that an AI's behavior remains consistent with human values and intent?
- Net speed
- Cloud link
- Goal sync
- Data flow
Explanation: AI Alignment (or Goal Synchronization) is the field of research dedicated to making AI systems act in accordance with human objectives.
What describes the 'Dark Patterns' in AI-driven web design that trick users into unintended actions?
- Fast access
- Bright modes
- Tricky design
- Hidden files
Explanation: Dark patterns are manipulative user interfaces that use cognitive biases to nudge users into making choices they didn't intend.
The annual 'AI Index Report,' which tracks global trends in artificial intelligence, is published by:
- Oxford Group
- Stanford Uni
- Harvard Team
- MIT Council
Explanation: The Stanford Institute for Human-Centered AI publishes the AI Index to provide a comprehensive look at the state of AI.
India's 'Bhashini' initiative is an AI platform primarily designed for which service sector?
- Translation
- Navigation
- Calculation
- Encryption
Explanation: Digital India Bhashini provides AI-led language translation to help citizens access digital services in their native tongues.
Which technology allows AI processing to occur locally on a device rather than on a remote server?
- Main AI
- Core AI
- Edge AI
- Grid AI
Explanation: Edge AI processes data locally (on smartphones or IoT devices), reducing latency and increasing privacy.
In machine learning, the phenomenon where a model learns the training data too specifically is called:
- Cross-link
- Under-load
- Hyper-scan
- Overfitting
Explanation: Overfitting occurs when a model captures 'noise' in the training data, leading to poor performance on new, unseen datasets.
Which ethical dilemma is commonly used to discuss the decision-making of autonomous AI vehicles?
- Turing Dilemma
- Halting Debate
- Trolley Dilemma
- Prisoner Debate
Explanation: The Trolley Problem is a thought experiment used to debate how AI should prioritize lives in unavoidable accidents.
Which process is used to artificially increase the size of an AI training set by modifying existing data?
- Data selling
- Dataset boost
- Data lockout
- Data removal
Explanation: Data Augmentation involves creating variations of existing data (like flipping images) to improve model robustness.
The IndiaAI Mission, approved in 2024, has been allocated an initial financial outlay of approximately:
- 10,000 Cr
- 30,000 Cr
- 40,000 Cr
- 20,000 Cr
Explanation: The mission was approved with an outlay of Rs 10,371.92 crore to build a comprehensive AI ecosystem including compute infrastructure.
In autonomous surveillance drones, the AI technology used to identify targets is called:
- Power check
- Object search
- File check
- Sound check
Explanation: Computer vision enables 'Object Detection' (Search), allowing drones to identify and track specific subjects in real-time.
The EU AI Act (2024) is the world's first comprehensive regulation based on which framework?
- Risk-based
- Time-based
- Cost-based
- Debt-based
Explanation: The act categorizes AI applications by the level of risk they pose to safety and fundamental rights.
In the context of AI development, what does the term 'Prompt' specifically refer to?
- System Speed
- Input Signal
- Storage Log
- Output Task
Explanation: A prompt is the initial text or command provided by a human user to guide the AI's generated response.
In 2024, the Indian Cabinet approved a flagship AI mission with an outlay of over:
- 10,000 Crores
- 50,000 Crores
- 5,000 Crores
- 1,000 Crores
Explanation: The IndiaAI Mission, with an outlay of Rs 10,372 crore, aims to establish a high-end computing ecosystem.
Which international organization issued the 'Ethics of AI' recommendation adopted by 193 member states?
- Ethics norms
- Trade norms
- Space norms
- Debt norms
Explanation: UNESCO issued the first global standard on AI ethics, focusing on human rights and dignity in AI development.
What is the technical definition of 'Zero-shot Learning' in the context of advanced AI models?
- Fast computing
- Cloud storage
- Task sans data
- System update
Explanation: Zero-shot learning allows an AI to perform a task it has never specifically been trained on, relying on generalized knowledge.
India's 'National Strategy for Artificial Intelligence' (NITI Aayog) is titled under the theme:
- AI for Ten
- AI for All
- AI for One
- AI for Now
Explanation: The #AIforAll strategy focuses on inclusive development and leveraging AI for social impact sectors like Health and Agriculture.
The 'Agriculture Stack' initiative in India uses AI primarily to provide help to which group?
- Farming aid
- Travel aid
- Mining aid
- Banking aid
Explanation: The AgriStack is a collection of digital records and AI tools designed to provide farmers with data-driven agricultural insights.
What is the primary purpose of 'Red Teaming' in the development of Large Language Models?
- Testing Speed
- Testing Reach
- Testing Safety
- Testing Sales
Explanation: Red Teaming involves adversarial testing where experts try to provoke the AI into outputting harmful or biased content.
India's 'Digital India Bhashini' initiative is an AI-led platform specifically for:
- Encryption
- Accounting
- Navigation
- Translation
Explanation: Bhashini aims to provide real-time translation across Indian languages to bridge the digital divide.
Which machine learning approach involves an agent learning through a sequence of 'rewards' and 'penalties'?
- Supervised
- Associative
- Clustering
- Reinforcement
Explanation: Reinforcement learning optimizes an agent's behavior by maximizing a cumulative reward signal from its environment.
Which nation held the Chairmanship of the Global Partnership on Artificial Intelligence (GPAI) for the year 2024?
Explanation: India chaired the GPAI in 2024, emphasizing a consensus on sovereign AI and collaborative safety frameworks.
Tamil Nadu's dedicated AI policy (2020) was one of the first in India to regulate state-run:
- State budget
- State taxes
- State export
- State ethics
Explanation: The policy provided a framework for the ethical use of AI and data in public services and government decision-making.
A 'Multimodal' AI model, such as Google Gemini 1.5, is characterized by its ability to process:
- Single input
- Logic blocks
- Mixed inputs
- Static data
Explanation: Multimodal AI can understand and generate content across different formats like text, audio, images, and video simultaneously.
What is the primary technical function of a GPU (Graphics Processing Unit) in the training of AI?
- Parallel ops
- File saving
- Single tasks
- Audio input
Explanation: GPUs are specialized for parallel processing, allowing them to perform thousands of mathematical calculations simultaneously for AI training.
The advanced technique of 'Few-shot Prompting' involves providing an AI model with a few:
- Random text
- Hint examples
- Single digit
- Simple query
Explanation: Few-shot prompting provides the model with a small number of examples to help it understand the desired task pattern.
The 'New Delhi Declaration' at the 2023 GPAI Summit emphasized the need for:
Explanation: The declaration focused on building trust and consensus for a global framework on sovereign AI and safety.
The 'Anthropic' safety policy for AI development is focused primarily on the principle of:
- Global reach
- Safety first
- Profit focus
- Speed levels
Explanation: AI safety companies like Anthropic focus on 'Constitutional AI' to ensure models remain helpful, harmless, and honest.
Which class of generative models is primarily used to create realistic images by removing 'noise' from a pattern?
- Code writing
- Image synthesis
- Text analysis
- File sorting
Explanation: Diffusion models generate images by learning to reverse a process that adds random noise to visual data.
What technical term refers to the internal variables an AI model adjusts during its training phase?
- System clocks
- Battery life
- Model weights
- Screen sizes
Explanation: Parameters (or weights) are the learned values within the neural network that determine how input data is transformed into output.
Which AI concept mimics the biological layers of the human brain to process information?
- Logic Nets
- Neural Nets
- Binary Map
- Linear Map
Explanation: Artificial Neural Networks (ANNs) are layers of nodes that simulate the way human neurons process data patterns.
The 'Asilomar AI Principles' (2017) serve as an international guideline for which sector?
- Ethics guide
- Law roadmap
- Space guide
- Sales guide
Explanation: These 23 principles cover research, ethics, and values to ensure AI development remains beneficial to humanity.
What describes 'Artificial General Intelligence' (AGI) which remains a theoretical goal in the field?
- Fixed-scale
- Logic-bound
- Task-driven
- Human-level
Explanation: AGI refers to a machine with the ability to solve any intellectual task that a human being can.
The 'Seoul Declaration' (2024) followed the Bletchley process to specifically address international:
- AI Trends
- AI Assets
- AI Safety
- AI Market
Explanation: The Seoul Summit focused on safety, innovation, and inclusivity, building on the safety benchmarks established at Bletchley Park.
What is the primary role of 'Computer Vision' in modern Artificial Intelligence applications?
- Text Analysis
- Audio Analysis
- Code Analysis
- Visual Analysis
Explanation: Computer vision enables AI to see and interpret visual information from the world, used in drones and medical imaging.
The 'Seoul Declaration' (2024) was an international agreement focusing primarily on high-level:
- AI Safety
- AI Bonds
- AI Sales
- AI Parks
Explanation: The Seoul Summit expanded on the Bletchley process, emphasizing safety, innovation, and inclusivity in AI governance.
Which type of learning involves a model being trained on a 'labeled' dataset with clear inputs and outputs?
- Clustering
- Predictive
- Supervised
- Generative
Explanation: Supervised learning uses a training set of labeled data to teach models to predict outcomes correctly.
What is the primary function of 'Generative Adversarial Networks' (GANs) in synthetic media production?
- Data Creation
- Data Sorting
- Data Storage
- Data Erasure
Explanation: GANs use two competing networks (Generator and Discriminator) to create highly realistic images, videos, and sounds.
The concept of 'Compute Governance' focuses on regulating which specific part of the AI supply chain?
- Storage bags
- Hardware caps
- Software ads
- Data exports
Explanation: Compute Governance aims to track and potentially limit the massive computing power (hardware) used to train large models.
The concept of 'Deepfakes' is most closely associated with which ethical concern in AI?
- Redundancies
- Misinformation
- Connectivity
- Inconsistency
Explanation: Deepfakes use GANs to create realistic but false audio/video, leading to significant risks of digital misinformation.
A designated AI 'Center of Excellence' (CoE) in India is typically focused on research in:
- Polar region
- Space travel
- Deep sea bed
- Medical apps
Explanation: CoEs are established in premier institutions to develop AI-driven solutions for sectors like healthcare and sustainable cities.
Which specific test, proposed in 1950, evaluates a machine's ability to exhibit behavior indistinguishable from a human?
- Hinton Test
- Turing Test
- Minsky Test
- Samuel Test
Explanation: Alan Turing proposed this test to assess machine intelligence based on conversational mimicry of human responses.
What technical term describes an AI model producing confident but factually incorrect or nonsensical outputs?
- Hallucination
- Misperception
- Fabrication
- Malformation
Explanation: Hallucination occurs when an AI generates text that is logically consistent in structure but factually wrong.
Which AI type is capable of performing a single, defined task like facial recognition or spam filtering?
- Global AI
- Strong AI
- Narrow AI
- Active AI
Explanation: Narrow AI (or Weak AI) is designed and trained for a specific task, lacking the broader cognitive abilities of humans.
Which Indian government portal serves as the 'National AI Portal' for all developmental updates?
- DIGI-AI
- BHARATai
- MEITYai
- INDIAai
Explanation: INDIAai is a joint initiative of MeitY and NASSCOM to serve as a central hub for the AI ecosystem in India.
The 'Technological Singularity' is a hypothetical future point where Artificial Intelligence reaches:
- Speed limits
- Data lossage
- Power outage
- IQ crossover
Explanation: The Singularity refers to a future point where AI intelligence surpasses human cognitive abilities, leading to rapid change.
What does the 'R' in the technical AI framework known as 'RAG' stand for?
- Relational
- Real-time
- Retrieval
- Recursion
Explanation: Retrieval-Augmented Generation (RAG) allows an AI to pull data from external verified sources to improve factual accuracy.
The landmark 'Bletchley Declaration' (2023) primarily addresses the international risks associated with:
- Narrow AI
- Legacy AI
- Static AI
- Frontier AI
Explanation: The declaration, signed by 28 countries including India, focuses on the safety of highly capable, multi-purpose AI models.
Which specific training phase involves adapting a general AI model to perform a narrow, custom task?
- Specific task
- Random guess
- System boot
- Generic flow
Explanation: Instruction tuning (or Fine-tuning) adapts a pre-trained model to follow specific instructions or perform in a specialized domain.
What is the primary risk associated with AI being labeled a 'Stochastic Parrot' by researchers?
- Infinite loop
- Low battery
- Meaningless
- Data piracy
Explanation: The term suggests that LLMs probabilistically link words together without any real understanding of the underlying meaning.
Which Indian technology startup developed 'Krutrim,' the country's first AI unicorn?
- Space probe
- Foreign app
- Indian LLM
- Global chip
Explanation: Krutrim, founded by Bhavish Aggarwal, is a Large Language Model designed specifically for Indian cultural and linguistic contexts.
In the context of AI safety, a 'Red Teaming' exercise is performed to identify vulnerabilities through:
- File search
- Hack testing
- Sales update
- Code cleanup
Explanation: Red teaming involves adversarial testing where experts intentionally try to provoke the AI into outputting harmful content.
In the context of generative AI, 'Tokenization' is the specific process of:
- Saving Code
- Buying Apps
- Selling Data
- Parsing Text
Explanation: Tokenization breaks down a string of text into smaller units (tokens) for an AI model to analyze.
India's fastest AI supercomputing system, currently ranked in the global Top 100, is known as:
- AIRAWAT
- MIHIR-A
- PARAM-S
- PRATY-S
Explanation: AIRAWAT is part of the National Program on AI and is installed at the Center for Development of Advanced Computing (C-DAC).
What is the primary objective of 'Explainable AI' (XAI) within the framework of digital governance?
- Transparency
- Connectivity
- Scalability
- Profitability
Explanation: XAI ensures that the processes and decisions made by AI systems can be understood and traced by human regulators.
The 'AI Elections Accord' (2024) signed by tech giants aims primarily to protect:
- Space treaty
- Tax reforms
- Online sales
- Voter safety
Explanation: The accord was signed to combat the deceptive use of AI-generated content (like deepfakes) in global election processes.
Which capability defines a 'Multimodal' AI model such as GPT-4o or Gemini 1.5?
- Audio-Only
- Text-Only
- Voice-Only
- Cross-Media
Explanation: Multimodal models can simultaneously process and generate outputs across text, audio, images, and video.
In the RLHF fine-tuning process, what is the 'H' that guides the AI's learning behavior?
- Static code
- Human input
- Random luck
- Fast timing
Explanation: Reinforcement Learning from Human Feedback (RLHF) uses human evaluation to train the model's reward function.
In the AI architecture known as 'RAG', what is the primary source of the information generated?
- Memory cache
- Power supply
- External data
- Internal code
Explanation: Retrieval-Augmented Generation (RAG) pulls information from external, reliable databases to ground the AI's response in facts.
The foundational neural network architecture used for modern Large Language Models (LLMs) is called a:
- Transceiver
- Transmitter
- Transformer
- Transductor
Explanation: Transformers utilize 'attention mechanisms' to process entire sequences of data simultaneously, enabling high-speed language tasks.