AI Fundamentals 2026: Safe AI Skills for Work

Explore what modern AI really is, from text and voice to images and data, and how it unlocks unstructured data to boost business productivity.

Define modern AI as more than a search engine, a knowledge work assistant that asks clarifying questions and generates drafts, plans, and checklists—while highlighting safety and judgment in its outputs.

Explore multimodal AI that fuses text, voice, and visuals to analyze documents and data. Prompts, tone matching, and source verification help ensure accurate outputs from spreadsheets, PDFs, and web results.

Explore how multimodal ai processes images, videos, and text, generate art and code, and empower nontechnical users to automate tasks with vibe coding.

Learn how the 80% unstructured data—dark data—drives AI value, as modern large language models unlock analysis of emails, contracts, and meeting notes without replacing structured data, delivering visibility.

Understand four AI integration patterns in business: enhance existing products, create AI native offerings, automate formerly manual tasks, and gain insights for smarter decisions.

Map AI use across three layers: outward facing systems, internal operations, and business intelligence. Prioritize testing, guardrails, and decision support to improve accuracy, security, and efficiency.

Discover how AI-assisted workflows boost productivity by accelerating research, analysis, drafting, and presentation tasks; save time, reduce costs, and enable teams to scale and stay competitive.

Use company-approved ai tools and follow your team's ai guidelines to ensure responsible handling of data. Avoid shadow ai by never using personal or free ai accounts for work tasks.

AI has exploded due to usability, accessibility, and affordability via natural language prompts. ChatGPT, Google, and Gemini now bring AI features to VS Code, Premiere Pro, and Photoshop.

Learn how large language models are probability-based pattern recognizers, why they hallucinate, and how generative ai raises copyright and plagiarism concerns while producing text, images, or code.

Explore how ai uses pattern recognition and probabilities to predict the next word and generate text or images, then compare it with human learning through touch, senses, and feedback.

Test ideas by running a simple prompt three times through ChachiPT to explore why a generative ai doesn't always give the same answer and compare results.

Uncover common ai misconceptions and fears by clarifying that ai recognizes patterns, not human understanding, and emphasize that modern tools are accessible to everyone, not just technicians.

Explore ai limits, including context size, token input, and lost in the middle; learn to verify outputs, guard against hallucinations, compare models, and apply prompt engineering for reliable knowledge bases.

Explore how AI hallucinates by predicting low-probability details and generating incorrect results, especially in images. Explain context, memory, prompt engineering, sentiment misreads, outdated knowledge, and privacy limits for safe use.

Explore the foundation of modern AI, including how it processes information, context windows, privacy, and hallucinations, and examine AI safety, responsible workplace use, and the EU AI Safety Act.

Explore how AI agents differ from large language models, with agents planning and acting to complete tasks—like querying data and sending email—while LLMs generate text and answers.

The lecture demonstrates using Claude’s Chrome plugin to act as an agent that searches London hotels for June dates, presents a plan, and guides booking up to checkout.

Showcase ChatGPT agent mode to book a hotel in London on June 3–6 via booking.com, with integrations to Google Drive, Gmail, and calendar.

The lecture demonstrates using AI agents to book calendar meetings and read Gmail messages, linking to Google Drive, scheduling prompts to function as your personal assistant.

Explore how an AI agent in Google AI Studio creates a plan to execute tasks with tools, builds code, outputs a web app, and contrasts with live language models.

Explore practical agent concepts by testing your language model's agent mode, browser plugin, and connections to schedule a meeting or book travel, while considering access levels and safety.

Explore responsible AI use in the workplace by examining risks of AI misuse, ethics, bias, accountability, and human oversight, while understanding copyright, intellectual property, and the EU AI Safety Act.

Discover how to avoid shadow AI risks by using AI safely and responsibly at work, protecting data and intellectual property while unlocking efficiency with large language models.

Identify major risks in AI misuse, including data breaches, reputational damage, security threats, copyright violations, and bias, with real-world case studies and guidance on testing for bias.

Explore real-world ai missteps and their consequences, from data leaks and misgoverned chatbots to bias and privacy concerns. Learn strategies for safe, governed ai use at work.

Explore a three-tier security framework for AI assistants: public tier for general use, enterprise tier with encryption and RAG for internal data, and on-premise tier for fully private, controlled environments.

Explore tier 2 enterprise-grade AI tools with zero data retention, encrypted communication, and GDPR compliance, designed to keep work data within a company-controlled ecosystem.

Practice spotting sensitive work data in prompts and selecting the right AI tool. Review ten workplace scenarios to decide between work, personal, or no AI tool, pausing to check context.

Maintain accountability for AI-assisted work by reviewing and validating outputs for accuracy, fairness, safety, and privacy. Watch for hallucinations and red flags, and escalate to a supervisor when unsure.

Audit a black box resume screening tool to reveal how training data and features like education, experience, hobbies, and zip code embed bias in hiring.

Explore levels of human oversight, from human in the loop with a reviewer for high-risk tasks to human on the loop, and finally no oversight for very low-risk work.

Define ownership of AI-generated content in the workplace, ensure use of company-approved tools and licensed data, and credit sources while avoiding copyright, licensing, and plagiarism risks.

Verify ai outputs against known facts, reliable sources, and company policies. Cross-check with multiple sources, assess bias and reasoning, demand evidence, and apply deployment due diligence with human review.

Explore risks of over-relying on AI, including loss of human skills and decision visibility, and assess continuity, explainability, and regulatory considerations like GDPR and the EU AI Safety Act.

Explore eu ai act and the four risk levels, prohibited, high, limited, and minimal, along with examples and the role of transparency for ai systems under ai safety act.

Regulators require high-risk AI to conduct risk management and keep a risk register. Maintain technical documentation, ensure meaningful human oversight, and test for accuracy and bias with evidence.

Describe the limited-risk AI category, where users must know when they interact with AI or see AI-generated content (including emotional recognition), with clear, understandable disclosures to preserve trust.

Apply data security best practices to minimal risk AI, reduce stored and processed data, and keep data accurate, secure, and up-to-date under GDPR. Respect privacy and consumer protection laws.

Explore AI safety basics and ethics, including misuse risks, three-tier security levels, bias, copyright, human-in-the-loop, and the EU AI Safety Act. Then learn practical prompt design to maximize AI outputs.

Learn to design effective prompts beyond basics with a three-step framework, a versatile template, and techniques like chain-of-thought and role prompting to shape AI outputs with purpose.

Learn prompt engineering, create detailed prompts for automation, and connect a custom GPT in an automation chain to scale AI work and explore AI agents with a practical use case.

Learn to tailor prompts by task complexity, provide clear outputs, and break complex requests into stages with company context, guiding the AI like a new hire.

Plan prompts by assessing task complexity and desired output. Draft, test, and refine prompts while evaluating risks, data safety, and copyright to achieve consistent one-shot results.

Explore a practical template for effective prompting, including role, task, limits, templates, and examples, and learn how these elements improve output through studies and practical tips.

Organize prompts by separating context, task, input, and output, using headings and delimiters to reduce ambiguity and encourage think mode processing for clearer, more accurate results.

Learn a simple prompt structure: assign a role, produce five key bullet points, and require using only provided information with an executive-ready tone.

Explore the core quadrants of prompts—task, role, limits, and guidance—to shape safe, accurate outputs. See how each element changes the output and the sources used.

Discover how emotion prompting boosts large language model performance in zero-shot tasks, with complex tasks improving by 115% and truthfulness and informativeness rising, plus practical prompt examples.

Boost accuracy with role prompting by assigning a role to the LLM, such as a data scientist or creative writer, and supplement with complementary descriptions.

Context provides AI with the environment, background information about your company, industry, and system to avoid generic results; craft prompts for an internal policy in a regulated financial services company.

Guide the live language model through complex tasks with steps and chain-of-thought prompting to improve accuracy, then demonstrate calculating liquidity and profitability, trends, and benchmarking for actionable insights.

Learn to use templates to standardize tasks, preserve structure, and scale outputs by guiding the language model's format and using examples.

Use good and bad examples to guide tone, format, and writing style through few-shot prompting, and place important information at the start and end of prompts to boost accuracy.

Master iterative prompting to improve AI output, refine one-shot prompts with templates and examples, and provide specific feedback to reduce vague, off-topic, or hallucinated results.

Learn to structure prompts, set the right context, and use roles and examples to shape outputs. Refine through iteration and explore these techniques in hands-on demo videos.

Apply prompt elements in practice by exploring meta prompting, role elements, pseudonymization, and context management, then adjust memory and model settings to observe output changes.

Master meta-prompting to build an automation with a custom GPT that analyzes dark, unstructured data, summarizes documents, and extracts key decisions and risks for healthcare operations.

Set up a GPT project prompt, organize prompts and outputs, and extract key decisions and risks from unstructured health care data in a PDF.

Explore how switching roles alters AI outputs for personal data analysis under UK GDPR, covering definitions of personal data, pseudonymization, and marketing risk implications.

Explain pseudonymization as a data security technique that scrambles sensitive information using a secure key table and cryptography, enabling masking with controlled access and later retrieval.

Attach the document once in the project folder, tweak the context, and run prompts to generate consistent outputs for a healthcare focused AI workflow.

Explore how chat history builds user context in memory for large language models, and how to manage memory in personalization and data controls to protect your information.

Discover how a structured prompt framework—defining task, role, limits, context, and format—improves AI outputs, with a comparison of vague versus detailed prompts to save time and enhance usefulness.

Review key prompt elements in action, from meta-prompting to memory settings, and see how they shape outputs, then explore personalized large language models, custom GBTs, and co-pilot agents for automation.

Learn to personalize large language models by creating custom GPTs and copilot agents, understanding prompts from tokens to output, and tailoring tools like Google Gemini Gems to your needs.

Explore custom GPTs, Copilot agents, and Gemini gems by configuring a name, description, an injected prompt, and a knowledge base to power enterprise tasks.

Explore custom gbt and Copilot agents, defining a stable persona with role, instructions, and reference material, and compare them with projects for long‑lived, consistent coaching, tutoring, or review tasks.

Set up a custom GPT called Document analyzer to read documents, produce concise summaries, extract key information and risks, and assess content sensitivity.

Build your first personalized ai assistant around real, repeat task. Set up a workspace in your ai tool, define its role, instructions, and style, and add references to save time.

Understand how prompts, instructions, and guardrails form a mega prompt. Learn how memory, knowledge bases, and retrieval augmented generation enrich outputs within the context window for safe, relevant results.

Explore tokens and the deep learning of large language models through vibe coding with Google AI Studio, showing how to build learning aids without coding using prompts and speech-to-text.

Explore how large language models process text by tokenizing into words or subword chunks, with examples from OpenAI's token platform and ChatGPT-5, noting token costs and evolving tokenization rules.

Explore how large language models predict the next word by computing word probabilities through neural nets and deep learning, adjusting weights and biases via supervised learning.

Combine prompts, system context, knowledge bases, and memory to guide a retrieval augmented generation, respect token and context window limits, and pass guardrail checks before producing output.

Shows how to use Google Gemini gems as custom GPTs with system prompts and knowledge bases to script YouTube videos, generating five concepts and viral titles from examples and transcripts.

Master personalized language models and configuring a custom GPT or Gemini gem, and explore vibe coding to build apps and websites with no-code tools like Google AI Studio and Repl.it.

Explore vibe coding with Google AI Studio and Repl.it, describe what you want to build, and watch AI create from scratch in real time.

Explore Google AI Studio, a developer-focused platform that lets you vibe code to build apps and web projects, adjust temperature, view token costs, and publish prototypes.

Demonstrates building apps with Google AI Studio, contrasts AI agents with large language models through planning and tool use, and showcases a bedtime story app with Gemini-powered outputs.

Explore Repl.it, a vibe coding platform that builds front end and database backends with integrations, and publish your app while weighing light and economy modes and costs.

Create a 16-bit lunar lander game in Replit, featuring real physics, automated testing with screenshots and self-correcting previews; explore checkpoints, publishing, and deployment.

Choose a simple repetitive task, use vibecoding to build a basic AI tool on a no-code platform, test it with sample data, and reflect on its use in your work.

Explore vibe coding with AI tools that turn ideas into working apps without traditional coding. Build automation chains and semi-automated workflows without coding, saving time on repetitive tasks.

Explore automation chains and semi-automation with NA10, connecting apps and triggering workflows without coding. Learn to use templates and marketplaces to install ready automations that run in the background.

Automate document analysis across folders with a prompt in n8n, producing a catalog of title, summary, key information, risks, and sensitivity, populated into a Google Sheet.

Monitor a drive folder for new or updated files, trigger automatic analysis with a large language model prompt, and append results to a spreadsheet catalog.

Build an automation flow with an agent using prompts, then debug with vibe coding and chat guidance. See how large language models aid troubleshooting and envisioning automation at work.

Explore the n8n marketplace and templates to quickly automate workflows with over 5,000 AI automations, including a ready-to-deploy Google Drive file organizer template that classifies documents, PDFs, and images.

Create an automation chain for a repetitive, multi-step workflow using AI tools, then map steps, choose a task, test the workflow, and document what it does and the time saved.

Wrap up this section by detailing workflow and AI-powered automation, showing how to trigger workflows, build agents, and use pre-built templates, with Gemini across Google Drive and Google Docs.

Explore how Google Gemini integrates across Google Workspace, including Docs, Sheets, Drive, Calendar, and Gmail, to act as a virtual assistant while understanding its work context limitations.

Explore Google Gemini's features, including a chat box, file uploads from Drive and Photos, deep research, image generation, and canvas editing, and customize with personal context and Gems.

Connect Google Drive to Google Workspace and access Google Mail, calendar, and Drive with Google Gemini; search documents, summarize a marketing strategy, and explore read-only limitations.

Read Google Drive, list files, and generate a content summary with risks and sensitivity; export to sheets is possible, but editing is not.

Learn Gemini's read-only limits with Google Drive, including handling PDFs and text but not images, plus safety-driven editing restrictions and how to enable in-document changes via the file itself.

Learn how Gemini and Copilot summarize, insert, and refine content in Google Docs and Sheets, including turning blocks into tables and placing a summary at the top.

Gemini Assistant in Google Sheets creates a name and summary table, summarizes documents from the marketing strategy, adds a target audience age, and generates bar charts to insert.

Explore creating and tweaking charts and pivot tables in Google Sheets with Gemini, from bar and line charts to pie charts, plus data analysis and trend insights with percentage changes.

Learn how Gemini reads Gmail, summarizes emails and Google Drive attachments, and assists with drafting replies, while keeping you in the loop and noting update limitations.

Discover how Google Gemini calendar lets you query your calendar and schedule meetings, including a 30-minute slot, while highlighting limitations like deletion, conflict handling, and read-only attachment visibility.

Explore Gemini across Gmail, Docs, Sheets, and Drive in a hands-on exercise. Select three tasks from different apps and assess time savings and output usefulness to build AI fluency.

Explore Google Gemini in action across Google Workspace, including Drive, Docs, Sheets, Gmail, and Calendar, recognizing current limitations and feature updates, and preview Microsoft Copilot across Word, Excel, and PowerPoint.