AI Governance Key Terms

LLM = A model with a large number of parameters, trained on enormous datasets, and used for text prediction. Machine learning = A subfield of AI where algorithms learn from data to make decisions or predictions. Machine learning model = A representation of underlying patterns in data created by applying AI to a training set. Misinformation = False content that is unintentionally misleading.

LLM = Can be used as generative or discriminative Machine learning = Implements algorithms that learn by experience Machine learning model = Used to make predictions on unseen data Misinformation = Can be spread via deepfakes

LLM = Its size is measured by the number of model parameters Machine learning = A problem-solving process with stages like data collection and testing Machine learning model = Created from a training dataset Misinformation = Can be created by those who lack intent to cause harm

<p>LLM = Predicts text based on learned word probabilities Machine learning = Used for applications like fraud detection Machine learning model = Learns from training data Misinformation = Lacks the intention to cause harm</p> Signup and view all the answers

<p>LLM = Falls under the umbrella of generative AI Machine learning = A subfield of AI Machine learning model = Learns patterns and relationships in data Misinformation = Lacks intent to harm</p> Signup and view all the answers

<p>LLM = Outputs text predictions Machine learning = Provides inferences and recommendations Machine learning model = Performs tasks on unseen data Misinformation = Results in misleading content</p> Signup and view all the answers

<p>Bootstrap aggregating = Machine learning technique Classification model = Machine learning algorithm Clustering = Unsupervised learning method Compute = Resource for AI tasks</p> Signup and view all the answers

<p>Conformity assessment = Evaluating compliance to standards Contestability = Allowing challenges to AI decisions Corpus = Collection of text data Human-in-the-loop = Involving humans in AI processes</p> Signup and view all the answers

<p>LLM = Processes word sequences Machine learning = Processes input data Machine learning model = Produces predictions or tasks Misinformation = Produces false or misleading content</p> Signup and view all the answers

<p>LLM = Learns from training data Machine learning = Learns iteratively from input data Machine learning model = Created by applying AI algorithms Misinformation = Created through means such as deepfakes</p> Signup and view all the answers

<p>Transformer model = A neural network architecture that learns context and maintains relationships between sequence data. Transparency = A broad term that implies openness, comprehensibility and accountability in the way AI algorithms function and make decisions.</p> Signup and view all the answers

<p>Hallucinations = Generated nonsensical content in AI models Impact assessment = Evaluating effects of AI deployment Inference = Applying a trained AI model to new data Interpretability = Understanding how an AI model makes decisions</p> Signup and view all the answers

<p>Input data = Data used as the initial point for machine learning Large language model = A deep learning model using very large amount of text or code Machine learning = AI approach enabling systems to learn from data Transfer learning model = Reusing of model from one task to another</p> Signup and view all the answers

<p>Transformer model = Leverages the technique of attention, i.e., focusing on the most important and relevant parts of the input sequence. Transparency = May refer to the extent to which information regarding an AI system is made available to stakeholders.</p> Signup and view all the answers

<p>Semi-supervised learning = Learning method which combines labeled and unlabeled data Small language models = Language models with fewer parameters Supervised learning = Learning from labeled data Synthetic data = Artificially generated data used for training ML models</p> Signup and view all the answers

<p>Transformer model = Comprehends the meaning of a word in the context of the whole sentence by attending to surrounding words. Transparency = Disclosing if AI is used through techniques like watermarking when dealing with language-based outputs.</p> Signup and view all the answers

<p>System card = Documents the performance of an AI system Testing data = Data used to evaluate an AI model's performance Training data = Data used to train an AI model Transformer model = Neural network architecture processing sequences</p> Signup and view all the answers

<p>Transformer model = Aids in improving model accuracy in sentence comprehension tasks. Transparency = Refers to maintenance of technical and nontechnical documentation across the AI lifecycle.</p> Signup and view all the answers

<p>Chatbot = Conversational AI interface Computer vision = AI systems that understand and interpret images Human-centric AI = Development focused on human values and needs Input data = Data used to create outputs</p> Signup and view all the answers

<p>Transformer model = Used in language learning tasks to improve model performance. Transparency = Explaining how the model works through model or system cards.</p> Signup and view all the answers

<p>Bootstrap aggregating = Method for improving prediction accuracy Corpus = Structured text collection for analysis or training Contestability = Ability to challenge decisions made by an AI Conformity assessment = Verification that an AI system meets expectations</p> Signup and view all the answers

<p>Transformer model = Uses attention mechanisms to better understand input sequences Transparency = Supports auditability of the AI system through documentation.</p> Signup and view all the answers

<p>Transformer model = Focuses on critical parts of textual input via attention. Transparency = In open-source context, refers to making the source code publicly accessible</p> Signup and view all the answers

<p>Transformer model = Enhances ability of AI to correlate words with one another in a given document Transparency = Ensuring AI system activities and outcomes are easy to comprehend.</p> Signup and view all the answers

<p>AI assurance = Includes conformity, impact, and risk assessments of AI systems. AI audit = Review and assessment of an AI system to ensure it operates as intended and complies with relevant laws. AI governance = System used to manage, oversee and regulate the development of AI technology Algorithm = Set of rules designed to perform a task or solve a particular problem</p> Signup and view all the answers

<p>AI audit = Focus is to identify risks and offer mitigation strategies in AI. Algorithm = Focus is the specific approach to solving a computational problem AI governance = Focus is the implementation, management, and oversight of AI. AI assurance = Focus is on ensuring AI systems are tested and compliant</p> Signup and view all the answers

<p>AI assurance = Activities include certifications and testing. AI audit = Involves identifying and mapping risks in an AI system. AI governance = Oversees the development, deployment and use of AI technology Artificial general intelligence = Possesses strong generalization capability to achieve goals</p> Signup and view all the answers

<p>AI governance = Helps manage risks to ensure responsible and ethical AI development. Algorithm = A set of rules to perform the task in AI. AI audit = It can map risks and offer mitigation strategies. Artificial general intelligence = Considered as theorethical field of research</p> Signup and view all the answers

<p>AI assurance = Ensures compliance with relevant standards through various mechanisms AI audit = Verifies the operation of an AI system against guidelines AI governance = Includes laws, policies and frameworks at various levels Artificial general intelligence = It is contrasted with narrow AI</p> Signup and view all the answers

<p>AI assurance = Ensures AI operates as intended and complies with relevant standards. AI audit = To identify risks and offer mitigation strategies AI governance = To regulate the use of AI towards objectives Algorithm = To perform a specific task</p> Signup and view all the answers

<p>AI assurance = A comprehensive system to evaluate AI system behavior and compliance AI audit = A process to check and ensure AI systems are operating correctly AI governance = Structures that manage AI technology Artificial general intelligence = The goal of developing AI to perform a variety of tasks at a human level</p> Signup and view all the answers

<p>Bias = Systematic prejudice in favor of a group. Automated decision-making = Process of making decisions by tech without humans. Bootstrap aggregating = Method that aggregates multiple model versions. Classification model = Model designed to sort input data into different categories.</p> Signup and view all the answers

<p>Artificial intelligence = Field of simulating intelligent systems. Chatbot = Uses natural language processing. Bootstrap aggregating = Making a model more stable. Classification model = Sorts data into classes.</p> Signup and view all the answers

<p>Bootstrap aggregating = Improves model stability using subset of dataset Classification model = Sorts data into categories. Bias = Systematic deviation from the truth. Automated decision-making = Decision making without human involvement.</p> Signup and view all the answers

<p>Artificial intelligence = May include automated decision-making. Automated decision-making = Process that uses tech rather than humans in process. Bootstrap aggregating = Method that strengthens model accuracy. Classification model = Used to sort data.</p> Signup and view all the answers

<p>Bootstrap aggregating = Can improve stability and accuracy of the model. Bias = Can be sourced through selection of data. Classification model = Takes input and sorts into different classes. Chatbot = Uses natural language processing.</p> Signup and view all the answers

<p>Bias = Can lead to discrimination. Automated decision-making = Uses technology to make decisions. Chatbot = Responds to text or speech. Artificial intelligence = Includes machine learning.</p> Signup and view all the answers

<p>Reinforcement learning = Training an algorithm to play a video game by trial and error. Reliability = Consistency in performance, making a system predictable. Robotics = Programming robot arms in an assembly line. Robustness = Maintaining accurate function with changed inputs</p> Signup and view all the answers

<p>Reinforcement learning = Optimizing decision making in dynamic environments. Reliability = Ensuring accurate prediction. Robotics = Physical interaction of AI with the real world. Robustness = Defending systems against unexpected attacks.</p> Signup and view all the answers

<p>Reinforcement learning = Utilizes feedback mechanisms of rewards and penalties. Reliability = Consistent and accurate system performance. Robotics = Combines AI systems and software to interact with the physical world. Robustness = Ability to withstand security threats</p> Signup and view all the answers

<p>Reinforcement learning = Improving decision-making through trial-and-error. Reliability = Improving prediction on unseen data. Robotics = Improving AI interaction with the physical world. Robustness = Enhancing an AI system's security and resilience.</p> Signup and view all the answers

<p>Reinforcement learning = AI learns through interactive experiences. Reliability = Ensures accuracy and consistency. Robotics = Enables AI to have a physical presence. Robustness = Protects against security breaches.</p> Signup and view all the answers

<p>Reinforcement learning = Training an autonomous vehicle. Reliability = Ensuring output is accurate and works as intended. Robotics = Automation in manufacturing plants. Robustness = Resilience and security of an online transaction system.</p> Signup and view all the answers

<p>Reinforcement learning = AI learns to achieve a defined objective guided by feedback. Reliability = Ensuring an AI system consistently performs its function. Robotics = Enables AI interaction with physical devices. Robustness = Making AI systems secure against attacks or failures</p> Signup and view all the answers