Lesson 1
focusing on the important information and ignoring irrelevant details
abstraction
an AI system that can process information from multiple types of input
multimodal model
the structured or unstructured information, such as text, images, or code, used to train an AI model and shape its ability to recognize patterns and generate accurate responses
training data
when a model favors certain information over others
AI bias
the act of giving credit to the original creator of a work, especially in the context of using or referencing their material
attribution
a question, instruction, scenario, or statement provided by the user to guide the AI's response
prompt
an interconnected network that makes decisions using weights and hidden layers. It is used to represent the relationship between words
neural network
when the model makes up false or misleading information that sounds real but isn’t true
hallucination
ensuring an AI system’s goals and outputs match human values, fairness, and safety
alignment
the legal right granted to the creator of an original work to use, distribute, and modify it
copyright
a technology that mimics human intelligence, performing tasks such as understanding language, recognizing patterns, and making decisions
artificial intelligence (AI)
a general purpose form of AI that is trained on a very, very large body of knowledge
large language model (LLM)
additional information or data surrounding a prompt or query that helps the AI generate more accurate and relevant responses
context
the process of creating precise and effective prompts to guide generative AI models
prompt engineering
a guide for responsible choices about emerging technologies, ensuring privacy, consent, and safety for everyone
ethics
The likelihood that a specific outcome might occur
Probability
the process of teaching an AI model by exposing it to large amounts of data, allowing it to learn patterns, improve accuracy, and refine its predictions over time
training
when an AI generates conflicting or opposing responses based on a similar input, making it impossible for both responses to be correct at the same time
contradiction
humans and AI working together, with humans providing creativity and direction, while AI assists by processing data, generating suggestions, or automating tasks
co-creating
the practice of openly disclosing how something works, especially in terms of decision-making, data use, or AI processes
transparency
What is the best way to improve the accuracy of an AI's response?
Provide a more detailed and specific prompt
How do multimodal AI models differ from traditional AI models?
Multimodal AI models can analyze multiple types of inputs like text, images, and audio.
Why do AI-generated responses sometimes seem wrong or off-topic?
AI uses randomness and probability, which can lead to unexpected results.
How does AI bias emerge from training data?
AI learns from patterns in training data, which may reflect societal biases.
Why might some people argue that AI developers should compensate original creators?
Because AI-generated content can be sold or monetized even when based on others' work.