Exam 3- AI

cognitive computing- observation

1.Observing behaviors, traits, occurrences, and bodies of evidence

Caktus AI

The first artificial intelligence (AI) software designed specifically for education.

output layer

The information from previous layers is analyzed.

ethical concerns surrounding AI

•privacy and surveillance •bias and discrimination •the role of human judgment

historical data

•represents data accumulated over time and stored for later examination. •It is an element in artificial intelligence (AI) and machine learning models. •By utilizing data, AI systems are trained to pinpoint patterns, trends, and correlations that might evade observation.

data scientists

•work with data to optimize product/service development and business strategies. •They develop data models and algorithms that are designed to investigate and automate processes and apply data mining and analysis techniques in model development and deployment.

What are the 4 things the FTC urges companies to do when they implement AI

1. be transparent wit consumers 2. explain how algorithms make decisions 3. ensure that decisions are fair robust, and empirically sound 4, hold themslevs accountable for compliance, ethics, fairness

6 steps of Deep learning

1.Collect data: In order for deep learning to occur, data must be collected. 2.Choose and optimize the algorithm: There are a number of different algorithms that can be used, and the choice is often determined by the task to be performed. 3.Setup and manage the environment for training: Many AI providers offer tools to assist in creating and managing the training process. 4.Train, retrain, and tune the models: As the process of analysis occurs, continual adjustments and tuning of process and input data should be executed. 5.Deploy models in production: This requires careful analysis in order for successful implementation to occur. 6.Scale and manage the production environment: Three important considerations are modularity, a multi-tiered approach to system design, and the ability to choose from a variety of deep learning frameworks.

CC- interpretation

1.Drawing conclusions from observable behaviors and past experiences to generate hypotheses about the meaning and possible courses of action

False Representation

: Information provided by AI chatbots should be treated in the same way as information from a journal article or book to ensure you are not committing plagiarism.

input layer

Data and information from a variety of sources is entered into the neural network through the input layer.

Chatbot Plagiarism

Research has shown that responses provided by AI chatbots are not always produced with unique responses. A response may contain information from a preexisting source.

Data Spillover

The use of data from individuals who were not the original targets of data collection.

data security

This includes taking proper security measures to protect all types of data.

hidden layer

This processes data sent by the input layer or from other hidden layers.

CC-decison

Using all the data gained from the previous steps to decide the best course of action or which decision to make

misinformation

While AI chatbots offer a wide range of instructional benefits, there are ethical considerations students must consider when using the software.

cloud computing

a computing model where processing, storage, software applications, and more are provided over a network, usually the Internet. Individuals access "clouds" of resources on devices connected to the Internet on an as-needed basis. Data, files, and software do not need to be stored directly on a device. Cloud computing and AI have become increasingly intertwined. Cloud-based AI applications: •The Internet of Things (IoT): designed to capture and process this large volume of data. •AI as a Service (AaaS): AaaS is one type of SaaS application. •Chatbots: computer programs that process and simulate human conversations.

Chat gtp

an AI chatbot auto-generative system developed by OpenAI. •An extensive language model trained on vast amounts of text data that allow it to learn patterns and relationships found in the data. •The primary function is to generate human-simulated responses to user queries.

Generative adversarial networks (GAN's)

are a type of machine learning model that uses two neural networks. networks are deigned to compete against each other to create artificial instances of data that are interpreted as real data

ML- reinforcement learning

covers decision-making processes. •Focuses on using machine learning to analyze a scenario which in turn creates learning about how a behavior can be optimized to achieve the best outcome. •Optimized behaviors are learned through extensive observation and analysis.

classification model

designed when machine learning allocates a label value to a specific class and then seeks to recognize these values to decide what categories they fit into •problems could include: •classifying an email message as phishing or not •classifying customer service issues

cognitive bias

includes conscious or unconscious errors in cognition that impact an individual's judgments, assumptions, and decision-making.

3 pos outcomes of AI

increased productivity, decreased human risk and injury, diminished human causes eorrors

AI detection software

interpret ext and potentially detect if AI generated the text

neural network

is a method in artificial intelligence that teaches computers to process data in a way that is inspired by the human brain. Neural networks are a type of machine learning process called deep learning. A neural network uses interconnected nodes or neurons in a layered structure that resembles the human brain. The neurons in a neural network are connected in three layers: •input •hidden output

unsupervised learning (ML)

machine learning algorithms are used to examine and cluster unlabeled data sets.

4 steps to cognitive computing

observation, interpretation, evaluation, decision

Natural Language Processing (NLP)

the study and application of programming techniques that allow computers to understans spoken words and text imputed by humans.

AI Engineer

use a variety of machine learning techniques, such as natural language processing and neural networks, to build AI models that are used to in AI systems.

deep learning

uses data and algorithms to create learning. But portions of the learning process are automated, eliminating some of the need for human intervention

Machine Learning

uses data and algorithms to emulate the way humans learn •relies on human interaction and input to assist in creating learning.

Supervised learning (ML)

uses data class labels within datasets that specify what the data represents. •Class labels are used to create the classification model.

User experince

•) focus on working with products and services to help users better understand their functions. •UX designers can incorporate AI to analyze customer data and to optimize product and service design. •The median annual salary for a job in user experience is $77,000.

AI majors

•Artificial Intelligence •Computer Science •Data Science Statistics

Guiding Principles for ethical use of AI

•Design AI technologies that are socially beneficial •Avoid forming or reinforcing unfair bias in AI systems •Follow laws, rules, and standards covering AI •Incorporate transparency and accountability concerning AI practices •Implement data security and privacy principles

generative AI

•Designed to generate new human-like content that is similar to the data it was trained on •Are not trained on any specific rules or constructs •Data used to train is often collected from the internet •Output is surfaced to the user via chat interfaces that are user-friendly and scalable to mobile apps.

Traditional AI

•Designed to perform specific tasks and to predict outcomes •Tasks are designed using rules, decision trees, and machine learning algorithms. •Data used to train is curated from data that is/was collected for a particular purpose •Output is surfaced to the user via dedicated applications, including BI reports and dashboards.

incomplete data bias

•If data is incomplete, it may not be representative of the population being measured, which can lead to bias. •It is important to analyze the design and implementation of AI systems to ensure bias is avoided.

IBM ladder approach to AI

•Infuse: Operationalize AI throughout the business. •Analyze: Collect and organize data in meaningful and reliable processes. •Organize: Create a business-ready analytics foundation. •Collect: Make data simple and accessible. Data collection should encompass a wide variety of data from across the organization

steps in building an AI model

•Problem statement and goal identification •Preparing and gathering the data •Choosing the algorithm •Analysis of Data •Training the model •Assessment and deployment •Monitoring, evaluation, and improvement

3 ways to build trust in AI platforms

•Transparency: Creating output that can be shared with the public about how systems use data and interact with populations should be available as well as education about AI systems and processes. •Human values: Instilling human values in AI can help to build trust. There is the risk that the AI programmers' biases and morals can play a part in determining AI outcomes. Creating strategies to mitigate the risk of bias and sharing these strategies publicly should be considered. •Collaboration: A multidisciplinary effort to advance AI is needed. Authorities across many disciplines need to collaborate to establish a common set of standards and platforms.

AI chatbot

•a computer program that incorporates artificial intelligence (AI) and natural language processing (NLP) to interpret user-provided questions and provide automated responses, thus simulating human conversation. •These programs use deep learning to enable more accurate responses over time. -chat gtp, caktus AI, AI powered bing

Structured Data

•organized in a manner that enables algorithms used in AI to search through it. •Commonly found in databases, which are structured in tables, rows, and columns. •Frequently utilized in AI applications that rely on defined data points for forecasting and inventory management tasks.

Unstructured data

•refers to information that lacks a predefined structure. •The majority of generated data falls into this category. •Analyzing data and deriving insights from it often involves employing AI techniques. •Unstructured information is used in sentiment analysis and in self-driving cars.

data repurposing

- the use of data outside the scope of the original purpose of the data

AI in health care

-robotic surgery -health tracking -pharmaceuticals

AI in business

-sales and marketing -accounting -IT oppurations - manufacturing

AI in governemnt

-automation of routine tasks -military strategy and warfare -policing

CC- evaluation

Determining which hypothesis makes the most sense

AI Language Models

computer software programs that use AI to process and generate human languages

drawbacks of AI

job losses, privacy concerns

cognitive computing and AI

seeks to better understand unstructured data. Based on human approach to decision making

external data

•comprises information sourced from entities outside the organization. •This data type is sourced from outside areas, including websites, social media platforms, government entities, and competitors. It has a range of applications.

machine learning engineer

•includes researching, designing, and testing machine learning (ML) models and systems as well as assisting in the training and retraining of ML models. Statistical analysis is used to analyze and improve ML models.

Information Architecture

•is about helping people understand their surroundings and allowing them to find what they're looking for in the real world and online.

internal data

•is information gathered within an organization. •Departments such as human resources, operations, finance, and procurement collect data that covers various operational aspects.