For use of Generative Artificial Intelligence (GenAI) and Large Language Model (LLM) Tools
| Audience | LCL employees, contractors, and third parties using AI technologies for Leah Chang Learning (LCL) |
| Publishing Location | LCL website |
| Used for/when | Project proposals, contracts between LCL and clients, and general discussions. |
This is a real-time policy that will and needs to evolve frequently.
Last updated: October 1, 2024
Sure, everyone is using ChatGPT. Here at LCL, we use it intentionally, and with client permission.
We are intentional and careful about our use of GenAI. Together, we can decide on an ethical use case, to scale, and make your project more efficient.
We’re all about making learning work smarter, not harder. That’s why we use the latest AI technology to bring efficiencies to how we as humans design learning experiences. We use paid AI tools to help us personalise learning paths, tweak content delivery for better results, and always review GenAI outputs with humans. Our principles include:
- Ask permission (respect IP)
- Prefer paid vs. free,
- Use vetted and sound (adult learning science) custom bots
- Educate ourselves and our project-clients
- Always reviewed by humans
- Stay wary (consult with SMEs)
- Continue challenging datasets
- Stay aligned as a team with these things
- Disclose use of AI when working on projects
LCL’s GenAi Philosophy
At Leah Chang Learning, we are a learning technology consultancy that approaches new tools with both curiosity and caution. We understand that generative artificial intelligence (GenAI) can simulate human intelligence to create answers, products, or perform tasks. However, it does not replace the essential role of human critical thinking needed to use these outputs in relevant, appropriate, and ethical ways.
We see AI as a technology that should be accessible to everyone, and with the potential to be without bias.
LCL use cases for ai:
- Accessibility: Rewrite, level of language, plain language
- Strategic Use: We use AI as a strategic tool to solve specific problems and to enhance the capabilities of our team and our clients.
- Automation: AI is a tool to help with task automation and scaling operations, enhancing human abilities without replacing jobs.
We as humans can improve GenAI data sets=> We aim to continually improve AI by providing it with accurate, human-reviewed data and content. We may encourage clients to intentionally input/upload true/vetted content into datasets for public use when/if appropriate.
Ethical Use of AI:
We are committed to using AI tools responsibly to avoid causing harm, discrimination, or unethical outcomes. This commitment includes:
- Ongoing professional development.
- Regularly questioning and reviewing our AI use.
- Respecting the choice of individuals and organizations who prefer not to use AI.
- Always seeking permission in advance for the use of AI.
- Being transparent about where, how, and when AI is used in our projects.
GenAI tools LCL uses
- Current Main Tools: We primarily use ChatGPT4, which is a paid account, with limited use of ChatGPT3, a free account.
- Specific Instructional Tools: We use AI agents designed for instructional design, such as TonyBot and FreeGeeks, available on both ChatGPT3 and ChatGPT4.
- Potential Future Tools: We are open to incorporating third-party or publicly available AI tools, such as Google’s Bard, Microsoft Pilot, and image generators like DALL-E 2 and Midjourney. We will notify you in advance if we decide to use any new tools.
How LCL uses AI
- Preference for Paid Tools: We choose paid AI services instead of free ones because they offer better security and features.
- Client Consent: We always get permission from our clients before we upload any of their materials or intellectual property into our AI tools.
- Accuracy and Alignment: We make sure that all AI-generated content is accurate and matches our company values.
- Legal Compliance: We follow all relevant laws and regulations, particularly those related to data protection and privacy.
- Responsible Use: We use AI carefully to prevent harm, discrimination, or any unethical outcomes.
- Data Care: We handle all data carefully, as if it could be widely shared. Only anonymized and non-sensitive data will be used to protect user privacy.
Examples of LCL use of GenAI
In all cases below, LCL will have human reviewers reviewing and editing any GenAI results/content. We don’t use it as is.
| LCL use case | Example | Information/data uploaded or accessed by GenAi | Decisions/approvals LCL clients can make with us |
| Operational – Meeting Notes | Summarizing meeting notes, action items | Meeting transcripts with names and timestamps | Permission to recordPermission to use AI transcription |
| Learning objectives validation or pre-development | Generating draft learning objectives or sample scenarios for eLearning modules | Outline of course topics and learning objectives | Permission to process provided course content |
| Quality Assurance – Writing Style | Ensuring consistency in tone and style across documents in line with client’s writing style guide | Writing style guide and course content draft | Consent to use writing style guide and content for AI analysis |
| Accessibility Compliance | Checking digital content for accessibility compliance | Digital content files (e.g. HTML, PDF) | Permission to assess content for accessibility |
| Translation | Translating content between English and French | Course materials in the source language | Permission to use course materials for translation |
| Research and Data Analysis | Analyzing data trends to inform content development | Aggregated data from learning assessments and feedback | Permission to analyze aggregated data |
| Custom AI model for Interactive Learning | Develop a custom AI model to simulate real-world scenarios relevant to the course material | Learner inputs from scenario-based exercises, feedback on performance | Permission to develop and train a custom AI model using course-specific data |
| Risk Management Enhancement | Using AI to predict and mitigate project risks | Project scope documents, risks from previous projects | Permission to analyze project documentation for risk assessment |
Examples of when NOT to use of GenAI
| LCL use case | Example | Rationale |
| Critical thinking | Analyzing survey responses in detail | |
| Credible source search | Citations/sources (hallucinations) | ChatGPT will hallucinate and make up sources and citations that do not exist. ALWAYS double check. |
AI Training, Enforcement, and Updates
At Leah Chang Learning, we are committed to continuously enhancing our AI practices through regular training and educational programs. We actively encourage feedback and want our team to feel comfortable discussing any AI-related concerns.
Reporting Concerns: If you have any issues with our AI usage, please contact the LCL leadership team immediately at projects@leahchanglearning.com.
Enforcement of AI Policy: Failure to adhere to this policy may result in disciplinary measures, up to and including termination of employment or contracts.
Policy Review and Updates:
Updates or changes to the policy will be communicated clearly to all employees, and we will provide additional training as necessary.
We regularly review our AI policy every three months to ensure it aligns with the latest best practices and legal requirements.