Generative Artificial Intelligence

Define fundamental AI terms and concepts.

• Artificial Intelligence

• General vs. narrow AI

• Recent history/timeline

• Enabling technologies

• Machine Learning

• Algorithm vs. model

• Datasets

• Approaches (supervised vs. unsupervised)

• Deep Learning

• ANN

• Network parameters/weights

• Hyperparameters

• Applications

• NLP

• Computer vision

• Model checkpoints

Define generative AI terms and concepts.

• Generative AI

• Key organizations

• Key resources

• Generative AI modalities - Text ◦ Code ◦ Images ◦ Video ◦ Audio ◦ Multimodal

• Generative AI approaches - GAN ◦ VAE ◦ GPT ◦ Diffusion ◦ RLFH

• Prompting ◦ Prompt engineering ◦ In-context prompting

• Fine-tuning models

• API access

Generate text using AI.

• LLM

• Tokens

• Approaches ◦ Text generation ◦ Text completion ◦ Chatbots ◦ Speech to text ◦ Transcription ◦ Assistive AI with text outputs

• Prompt engineering for text generation

• Fine-tuning text models

• Common text generation tools ◦ OpenAI GPT models ◦ ChatGPT ◦ Bard ◦ LLaMA ◦ Microsoft 365 Copilot ◦ Duet AI ◦ Whisper

• Business use cases ◦ Customer service ◦ Online ordering systems ◦ Marketing campaigns ◦ Information summarization ◦ Information mining and inference ◦ Language translation ◦ Interviews and onboarding

• Generate code using AI.

• Generate images and video using AI.

• Generate audio using AI

Identify shortcomings of generative AI.

• Confabulation/hallucination

• Misinformation and misleading content

• Cost ◦ Monetary cost of implementation ◦ Training time cost ◦ Output time cost

• Lack of fine-tuned control

• Limitations of training data

• Hardware requirements

• Reliance on external factors ◦ Cloud services ◦ Foundational models

• Industry-specific risks

• Adversarial vulnerabilities

Identify ethical risks of generative AI.

• Privacy issues

• Accountability issues

• Transparency/explainability issues

• Bias/discrimination issues

• Safety/security issues

Identify business concerns of generative AI.

• Governance

• Employee impact

• Future of work

• Operational risks

• Data risks

• Brand reputation/consumer trust risks

• Legal issues ◦ Intellectual property/copyright ◦ AI laws and regulations

Apply best practices for implementing generative AI in the organization.

Evaluate the results of generative AI projects.