Generative AI Flashcards
Define Artificial Intelligence in the context of information technology.
AI attempts to replicate human or near-human intelligence in machines and encompasses machine learning and deep learning.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
How does Machine Learning relate to creating data models?
Machine Learning uses data and algorithms to create models, which are mathematical expressions approximating the relationship between variables in the data.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
What is Deep Learning and how does it differ from Machine Learning?
Deep Learning is a subset of Machine Learning based on neural networks, using connected layers to create artificial approximations of human neural networks.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
List three types of content that Generative AI can create.
Text
Visuals (Images, videos, 3d models, graphs)
Audio (music and speech)
———————————————————————
📚 Introduction to Generative AI (NowLearning)
How can Generative AI contribute to diversity and inclusivity?
By reducing bias and discrimination in areas such as healthcare, education, and employment. + Providing accessible and personalize solutions.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
Name two key business benefits of Generative AI.
It can increase opportunities in cost and productivity, and enhance revenue by creating intelligent, original content.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
What limitations and risks are associated with Generative AI?
- Difficulty adapting to new circumstances
- Exhibit harmful biases
- Pose intellectual property concerns
- Lack transparency
- Introduce cybersecurity dangers
- Sometimes it generates inaccurate content
———————————————————————
📚 Introduction to Generative AI (NowLearning)
How does ServiceNow use generative AI capabilities?
ServiceNow uses Generative AI for language generation to take text inputs from various sources and generate new content efficiently.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
What five generative AI capabilities are implemented out of the box?
- Generate
- Summarize
- Generic Prompt
- QnA
- Sentiment Analysis
———————————————————————
📚 Introduction to Generative AI (NowLearning)
What is the purpose of ServiceNow’s Now Assist?
Now Assist integrates generative AI features into the platform to enable automation, accelerate productivity, simplify repetitive tasks, increase agility, and transform user experience.
———————————————————————
📚 Introduction to Generative AI (NowLearning)
What is the purpose of the Generative AI Controller?
It integrates Generative AI functionality in ServiceNow workflows, allowing content generation through providers like Azure OpenAI and OpenAI.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are the essential capabilities of the Generative AI Controller?
It connects to AI providers, interacts with AI models, and offers a seamless design experience via Flow Designer, Virtual Agent, and Scripting on the Now Platform.
———————————————————————
📚 Now Assist Essentials (NowLearning)
How does the Generative AI Controller work?
It connects APIs to the Now Platform, sends requests to AI providers, which process and send back responses for various applications in ServiceNow.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are the ideal and non-ideal criteria for implementing Generative AI Controller?
Ideal
- Using Flow Designer or Virtual Agent
- Aiming to reduce content creation time
Non-Ideal
- Having a self-hosted instance
- Operating in a regulated market
- Using a mid-server for internet resources
- Not being mature with existing AI products
———————————————————————
📚 Now Assist Essentials (NowLearning)
Define “Prompt” in the context of Generative AI.
It’s the textual input supplied to the LLM to process and generate content. It can be a question, statement, or instruction.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What data is suggested for the Generate Content capability?
Records containing text, such as
- Knowledge Articles
- Ticket Descriptions
- Notes
- Work Notes
- Comments
———————————————————————
📚 Now Assist Essentials (NowLearning)
What characteristics should the data possess for accurate content generation?
The data should be high quality, error-free, diverse, relevant to the task, and requests to Generative AI should be precise.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are some supported customization for the Generative AI Controller?
It supports different ChatGPT models and integrates well with Flow Designer, Virtual Agent, and in custom scripts.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What APIs come out of the box with the Generative AI Controller?
- Completions
- Chat Completions
- Moderation APIs
———————————————————————
📚 Now Assist Essentials (NowLearning)
What type of instances are the Generative AI Controller not supported on?
- Self-Hosted
- GCC (GovtCommunityCloud)
- FedRamp (Federal Risk and Authorization Management Program)
- PDI
———————————————————————
📚 Now Assist Essentials (NowLearning)
What ServiceNow products are able to use the Generative AI Controller?
Yes, it is a platform capability that all applications can leverage.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are the prerequisites for installing and using the Generative AI Controller?
- ServiceNow instance must be on specified version
- Microsoft OpenAI or OpenAI subscription
- Alias Credentials (API Key)
- Now Assist Subscription
- Necessary Apps Downloaded
———————————————————————
📚 Now Assist Essentials (NowLearning)
What role does Integration Hub play in the Generative AI Controller?
It enables the execution of third-party APIs as part of a flow when a specific event occurs in ServiceNow.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are Generative Models and LLMS?
- Generative Models are AI Models that generate new data based on input.
- LLMS (Large Language Models) are generative AI models capable of generating human-like language at scale.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What is a hallucination in the context of Generative AI?
It’s when the LLM generates incorrect output due to reliance on its internal “knowledge”, which is not applicable to the user query.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What is the suggested data for the QnA Capability?
Knowledge Articles
———————————————————————
📚 Now Assist Essentials (NowLearning)
What does good-quality data for machine learning entail?
It should be diverse, high-quality, free of errors, and relevant to the task.
———————————————————————
📚 Now Assist Essentials (NowLearning)
What types of ChatGPT models does the Generative AI Controller support?
Legacy and newer models like gpt-4 and gpt-3.5 turbo
———————————————————————
📚 Now Assist Essentials (NowLearning)
What are the purposes of Completions, Chat Completions, and Moderations APIs?
- Completions and Chat Completions: Generate content using different ChatGPT models.
- Moderations: Checks content compliance with the vendor’s usage policies.
———————————————————————
📚 Now Assist Essentials (NowLearning)
