| Version | Date/Time | Release Notes/Update |
| v1 DRAFT | 2/29/24 @ 11:59 pm | A Personal Leap Ahead |
| v1.5 | 3/6/24 @ 11:31 am | Working draft with inputs from Paul Owen mk.marsden@touchpointstrategies.com nmcmichael@liveramp.com douglandman@gmail.com |
Introduction
Welcome to a concise guide on how I’m integrating Artificial Intelligence (AI) into my personal consulting practice.
At Perceptint, I’m committed to leveraging Artificial Intelligence (AI) to enhance my strategic digital strategy and marketing services, extend the core capabilities I can provide for my customers, and improve the operations and efficiency of my business. My aim is to utilize AI not just as a tool for innovation, but as a means to set new standards in ethical AI usage in the consulting industry. This document outlines my approach to using AI, including my objectives, use cases, ethical considerations, and governance.
As I mentioned earlier, this document started as a personal experiment for my business. In the process of research and asking my network for feedback, I found others who immediately recognized the hole I saw and shared the need for some guidelines. From Friday March 1st to today, March 7th, it’s evolved from a collaboration, to now approaching it as an open-source Manifesto.
As a small business, this is the first “formal guidelines” I’ve written for myself. With the current state of AI and the rapid advances happening daily in 2024, I felt it was vital to set a personal level of public accountability. I also want to set an example for what I expect most businesses to have in place within the next 12 to 18 months.
As updates are made, I’ll update the version number of the document. I’ll keep each version archived so you can see how it evolves.
My Goals using AI at Perceptint, LLC
Purpose
My goal is to experiment with building an AI assisted small company by integrating AI seamlessly into my operations, ensuring it aligns with the goals of the business. I believe it will assist me in various ways to automate processes and deliverables so I can continue improving my consulting services to a larger number of clients.
I also want to state that I know AI is not a fix or magic wand for everything. If anything, to get better from using AI tools, I have to put in more work up front.
Application
As a technologist, I’m trying to see if the promise of AI will live up to the current hype. I hope that AI will help me improve deliverables, assist in drafting document ideas for me to customize (including this set of Guidelines), for editing and brainstorming, data analysis, be my assistant in automating operations, enhancing my customer service, and offering additional resources like chatbots that extend my capacity. I preview and trial the newest technologies to provide the highest service value. These tools could help me deliver customized solutions and insights.
As I identify and implement additional use cases, I will update the policy to reflect the current list of the tools I use.
Data Management
Client confidentiality is paramount. I prioritize the privacy and security of my clients’ contact information. All AI applications and use of AI technologies will adhere to strict data management protocols, ensuring that any future data collection is used responsibly and ethically.
Data and the rules for data management for my customers always reside in their systems, not mine. As I add to the Guidelines, this section will grow to include the types of data and rules for the data I’m feeding into AI tools, where it lives, including the outputs, and how additional training and fine-tuning is taking place.
Ethical AI Use – My Commitment to Transparency
Responsible & Ethical AI use is at the heart of my operations. I commit to maintaining human oversight over all AI applications, ensuring accountability for AI outputs, and adhering to the highest ethical standards. I’m personally responsible for everything I deliver, including AI created materials.
AI Governance Structure – My Commitment to Open and Shared Evolution
As the principle and only decision-maker, I maintain rigorous oversight over AI initiatives. I’m creating an advisory group in 2024 to consult with a network of seasoned professionals to validate my AI strategies, ensuring they align with industry best practices.
Privacy & CyberSecurity Assurance
I’m committed to safeguarding client data with the latest security measures, ensuring compliance with evolving privacy laws. My AI applications will always be evaluated for their security and privacy safeguards. I will ensure my customers and myself go through annual training for privacy and security measures to ensure we mutually share in the responsibility. If resources aren’t in place before I start, I will work to bring partners to the table who can satisfy the agreed upon training needed.
As new AI applications or processes are tested, I’ll involve my customers to help them learn and adapt what I’m doing for them and expand upon it.
Stakeholder Engagement
Purpose:
To foster an inclusive, transparent dialogue with all stakeholders, including clients, partners, employees, and the broader community, regarding my AI initiatives. My goal is to ensure stakeholders are well-informed, have their voices heard, and are actively involved in shaping our AI journey together.
How I’ll do it:
-
- Drive the Dialogue: I will be active in advocating for stakeholder education and their participation in the bigger dialogue we’re starting.
-
- Feedback Channels: Establish open, accessible channels for stakeholders to share their insights, concerns, and suggestions about my AI use. This could include talking to chatbots on my behalf, SMS services, taking a quick survey, general feedback forms, or a dedicated email address. As these are created, they will be documented here.
-
- Educational Initiatives: Offer webinars, workshops, and resource materials to demystify AI technologies, outline my ethical AI practices, and discuss the benefits and challenges of AI in the innovation industry. I will actively promote AI literacy and training within the organizations I work with to ensure that all employees understand AI’s potential, limitations, and ethical use. My goal is fostering a culture of informed and responsible AI use.
-
- Publish use cases, best practices, and failures: As I have learnings and examples of how the Guidelines are implemented from my own use or the community, I will share them as resources back with the community. It’s especially important that we not treat this as “fail fast” and “release often” type of thinking.
Details matter.
Understanding what went wrong matters.
Advancing shared knowledge is critical for us to make advancements, not retreats.
- Publish use cases, best practices, and failures: As I have learnings and examples of how the Guidelines are implemented from my own use or the community, I will share them as resources back with the community. It’s especially important that we not treat this as “fail fast” and “release often” type of thinking.
Monitoring AI Performance & Evaluation
Purpose:
To ensure my AI applications are effective, ethical, and aligned with my business goals and stakeholders’ expectations. This involves continuous assessment of AI’s impact on my operations, client satisfaction, and adherence to my ethical principles. I will also look at ways to use AI agents to assist in monitoring my compliance with these guidelines where possible. Humans will always be involved at every step, ensuring we don’t automate our way into more unintended consequences in this exciting time.
Bias in data and opinion will always be examined and is expected to be part of the open dialogue. It’s about finding the errors in how we did business before, so we can make radical jumps ahead with how we do business in the future.
Categories of AI Being Tested:
-
- Chatbots/Conversational AI
-
- Example: Virtual assistants and customer service bots that understand natural language.
-
- Description: Interfaces that comprehend and respond in human language, offering support, information, or assistance.
-
- Chatbots/Conversational AI
-
- Text Prediction/Assistance
-
- Example: AI writing assistants for real-time suggestions and corrections.
-
- Description: Aids in writing and other text tasks by generating predictive text to enhance productivity and accuracy.
-
- Text Prediction/Assistance
-
- Classification
-
- Example: AI for categorizing images, documents, or web content.
-
- Description: Organizes data into categories, such as spam detection in emails or topic classification for articles.
-
- Classification
-
- Analysis/Predictive
-
- Example: Tools predicting market trends and consumer behavior.
-
- Description: Utilizes historical data for future event predictions, including sales forecasts and fraud detection.
-
- Analysis/Predictive
-
- Generative Text/Art/Video
-
- Example: AI creating articles, artwork, and videos from training data.
-
- Description: Generates new content, including digital art or video clips, based on textual descriptions.
-
- Generative Text/Art/Video
-
- Text-to-Video
-
- Example: Generative models creating video content from text.
-
- Description: Revolutionizes media production with AI-generated video content.
-
- Text-to-Video
-
- Business Process Automation (BPA)
-
- Example: Automated customer service, HR onboarding processes, and financial operations.
-
- Description: Utilizes AI to streamline and automate routine business processes, reducing manual effort and improving efficiency. This includes automating repetitive tasks, optimizing workflow, and enhancing decision-making through AI-driven analytics.
Specific AI powered tools I’m using or testing:
-
- OpenAI suite – ChatGPT, Dalle, GPT Builder, Bing, API
-
- Gemini Pro in Google Suite
-
- Grammarly – AI assistant
-
- LinkedIn – Premium (ChatGPT integration from MS)
-
- Zapier – Service connector
How I’ll Evaluate:
-
- Performance Metrics: Define clear, measurable goals for each AI application, such as improved client engagement, reduced operational costs, or enhanced decision-making accuracy. These metrics will guide my evaluation of AI’s effectiveness.
-
- Ethical Audits: Regularly review AI applications with my AI advisory group to identify any ethical concerns, biases, or unintended consequences that could occur. These audits will help me maintain my commitment to ethical AI use.
-
- Iterative Improvement: Use insights from performance metrics and ethical audits to refine and improve my AI applications. This includes updating AI models, adjusting data inputs, and revising use cases to better serve my stakeholders.
-
- Create New Measures: I believe that once businesses evolve into being AI powered where it fits we will inevitably have new business success measures, approaches to value creation, and ultimately business success.
-
- If it’s not working, recycle the idea and trash the AI: I have an obligation and right to stop and pivot out of solutions that don’t work. My expertise will be used finding applications for AI that work well and scale.
Additional Considerations
Addressing Data Bias, Inclusivity & Diversity:
I commit to deploying AI technologies that promote inclusivity and diversity. This includes using diverse data sets to train AI models, actively seeking to eliminate biases, and ensuring AI applications serve a wide range of needs and preferences.
Sustainability:
I recognize the environmental impact of AI technologies and commit to sustainable practices. This includes optimizing computing resources, selecting energy-efficient technologies, and supporting research into reducing AI’s carbon footprint.
Transparency & Explainability:
I will always advocate for transparency and explainability of AI decisions and processes in areas where that decision making intersects with customer interests or rights. The goal is to always enhance stakeholders’ trust. I understand that proprietary algorithms are something companies have a right to deploy. Whenever possible, detailing how the “black box” works – at least for what’s going in and going out, is critical for overall trust with these tools.
Legal & Regulatory Compliance:
I pledge to stay abreast of and comply with relevant legal and regulatory requirements affecting AI applications. This involves regular reviews of compliance frameworks, engaging with legal experts, and adapting my AI practices to meet evolving standards.
By writing this document and making it open-source, I commit to being a guide for the conversation around compliance, the thorny topics we’ll run into along the way, and being a sounding board for anyone interested in real, thoughtful dialogue and discussion. I’m a competitive cross examination debate inspired thinker, asking the world to take a stoic approach together towards AI. We have to.
Impact Assessment:
I am committed to assessing and documenting the impact of AI applications on stakeholders and the environment as part of a structured approach to evaluating benefits and risks.
Feedback & Continuous Improvement:
For these guidelines to have any long-term impact for a business, they must be updated at least once a year. Those updates should be based on new insights, technological advancements, or changes in legal requirements. And for every piece of feedback and input you receive along the way, show the audience where it fits (or doesn’t) in your next release.
Conclusion
I pledge to lead by example in the responsible use of AI and emerging technologies.
-
- By engaging internal and external stakeholders
-
- Through transparency in my data, toolsets, AI usage and outputs
-
- By considering additional ethical and environmental factors
I plan to use AI and evolving technologies to drive continuing innovation for myself and my customers. I also commit to use it to uphold my core values and human responsibilities to myself, my family, friends, society, and our planet.
Contact me if you have any questions about this policy or suggestions for future versions.
Content Authenticity Statement
70% of this Guidelines for Responsible AI & Business was written by me, Keith Boswell. My original intent with this was to publish for myself as a test to produce a working model quickly over two to three weeks. As it grew, I received constructive comments and feedback from my network.
I worked with a test version of an AI Policy Generator GPT to produce a draft outline and help me work through the questions I needed to answer for v1. This private GPT is being trained by myself and Paul Owen from Cyberdelic on the latest frameworks and news from sources like the NIST AI Risk Management Framework, the Responsible AI Institute, the EU’s looming AI ACT, and more. When the GPT is made available publicly, we’ll publish a link to all data sources.