Arpit patel - Artifact 6

6.5 Assignment: Personal Framework - Seeing Myself as a Change Leader

Machine Learning

Indiana Wesleyan University

APRIL 2025

Introduction

During this course, I learned a lot more about myself and how my skills and progress as an AI and ML leader could be critically analysed. Particularly in this fast-moving era of AI/ML integration, the lessons, principles, and practices that I’ve encountered have shaped and refined how I view leadership. I have built this personal framework documenting my mission, values, goals, plans, experiments, and learning as an AI/ML leader. I look forward to continuing my growth by learning to be a thoughtful, ethical, and successful leader in the AI/ML landscape by unpacking my gut and how my future will play out.

Mission Statement

I am an AI/ML leader, and I want to help individuals and organizations discover ways they can leverage AI/ML technology responsibly and intentionally in ways that inspire innovation, ethical use, and positive social value. My dream is to drive a course of lifelong learning, interdisciplinary collaboration, and ethical AI to develop transformative technologies that augment human potential and address complex global challenges. I intend to propel sustainable growth within my organization and the greater AI/ML community by embracing holistic approaches of inclusivity, transparency, and ethics with AI/ML initiatives.

Core Values

As an AI/ML leader, I will use the following tenets to make decisions and take actions:

Integrity: Integrity is the basis for fostering trust in AI systems. I will always prioritize ethical issues regarding transparency, fairness, and traceability in AI. I dedicate myself to exercising control over AI systems to protect against the violation of privacy, maintain strict accountability, and ensure decision-making transparency.

747-746 In this context, what can AI/ML do? Collaboration is important to me and I think the best solutions come from diverse mind sets working together. Whether with engineers, business leaders, or policymakers, collaboration will be my leadership approach.

Keep up with the trend: AI/ML is also a fast-growing field so is my goal of continuous learning. I will desire to keep abreast with the latest latest advancements, tools and methodologies to maintain an edge in my domain. I have learned to take experience from the past to avoid making the same mistake again.

Ethical Responsibility: AI/ML technologies shape society in ways that cannot be ignored. As a leader, I will put ethics first in my decision making, engaging in innovative pursuits without sacrificing human rights or the greater good of society. This is especially important as AI technologies are increasingly shaping policy, healthcare, finance and day to day life.

The first is empathy — a sense for the needs, wants, and challenges of those who may be impacted by AI/ML systems. I will lead with empathy, making sure we rollout tools that keep the user at the centre and empower everyone — particularly under-represented groups who may be sidelined.

Key Objectives

To complement my mission and core values, I created key objectives that are:

Deep Dive into AI/ML: My first goal is to keep upskilling myself on these AI/ML technologies to become an expert on advanced machine learning algorithms, neural networks and deep learning models. When it comes to studying and learning, I try to stay updated with the latest trends in the industry - specifically, RL, CV and NLP.

SMART Goal: By the end of 2025, complete at least two advanced online courses or certifications on the newest AI/ML technologies (e.g., reinforcement learning, generative adversarial networks)

Create a Norm of Ethical AI: One of my long-term goals is to build a culture of ethical AI in my organization or team. I will advocate for ethical AI practices to ensure the AI models we create are fair, transparent, and accountable and do not amplify social biases.

This could be a measurable goal on personal basis: Establish and enforce ethical AI development standards within my team, with regular audits for bias and fairness in AI systems, by mid-2026.

Promote Cross-Disciplinary Work: AI/ML has [Safi] interdisciplinary roots, and I hope to spearhead efforts that aggregate knowledge from diverse domains — tech, business, health care, law. With the ultimate goal of enabling teams to keep all parts of their incredible product in sync, I use my cross-discipline experience to be the voice translators between highly technical teams and stakeholders who may not speak the same dialect.

Measurable Goal: (Conduct at least 02 cross-functional to 03 projects in 2025 with different type of people (engineers/legal experts/business leaders, and people from different regions).

Action Plans

In order to meet these intended goals, I created the detailed action plans below:

Short-Term Plans (2025)

Further Steps: Upon data attaining till January 2022 I will ignore platforms like Coursera, edX, and some AI/ML academies.

Attend AI Conferences and Seminars: Attend AI/ML conferences like as NeurIPS, and ICML, as well as local industry seminars in order to broaden my network and keep ahead of emerging trends, as well as learn from thought leaders in the AI field.

Involve in AI Ethics Committees: Get involved with AI ethics committees or advisory boards, discussing how AI affects the society as a whole and working towards developing human-centered AI.

Long-Term Plans (2026 and Beyond)

Leadership Skills: Somethings like there — Leadership Development: To further develop my leadership skillsI will rely on mentorship from experienced AI/ML leadersand participate in leadership development initiatives that cover areas such as strategic decision-making, change management and innovation.

I will initiate the creation of a knowledge sharing platform in my company, where the employees are free to discuss the latest AI/ML technologies and discuss ethical challenges. This maintains our teams in sync and in the know.

Evaluation and Adaptation

AI/ML is a constantly evolving field, and as such, the initial implications are always evolving and must be revisited. These will be set as the mechanisms by which I will measure my progress:

Check-in every 6 Months: I would inspect myself against my goals every 6 M to keep me accountable. I will use this self-assessment to recognize skills I need to further develop and alter my goals accordingly.

Ask for feedback: Regularly, I will seek to collect informal feedback from my peers, other mentors and team members regarding my leadership, if I am creating a culture in our team where we are using ethical AI and building on each others ideas and input. Now I shall find feedback and develop it вследствие this and prepare accordingly.

Evolving with the Times: AI is constantly changing, and so will my framework. I will also continue to proactively seek out new industry trends, regulatory changes, and technological advancements that will demand adjustments to my tactics and goals.

Conclusion

This is a personal conceptual framework I have as my vision for what an effective, ethical, and innovative AI/ML leader would be. In doing so, I hope to make positive contributions in the field of AI/ML by adhering to my personal values, continuing to grow my technical and leadership abilities, and encouraging collaboration and ethical responsibility. This framework will not be stale though, you will be improving your plan through reflection, feedback, and adaptation, and you will revise your new baseline as you keep moving toward your vision that would make a dent in AI/ML.

6.2 Assignment: AI Incidents and Ethics

Machine Learning

Indiana Wesleyan University

APRIL 2025

Case Selection: Microsoft Tay Chatbot (2016)

Summary:

Example 1: Just in 2016 Microsoft released an AI chatbot called Tay that was designed to mimic human language based on what it learned from Twitter users. In only 16 hours, Tay had been manipulated by users with inappropriate comments, tweets and replies, resulting in highly offensive, racist and sexist tweets. The bot started encoding the data it was fed, with no safeguards to prevent learning bad behavior. Microsoft was forced to take Tay offline in a matter of hours, and the episode soon became a textbook case of how AI can be manipulated by bad data.

Critical Analysis

Problem 1: Missing Content Filtering Mechanisms

Why I chose this issue:

Tay’s case revealed the consequences of letting AI onto public spaces without controls. I chose this issue because content filtering should be a core capability in any generative AI exposed to users.

My thought process:

At first, I viewed this as a failure of public moderation. But as I explored further, I found that responsibility also falls on creators. Tay wasn’t repeating bad content — it was learning, and adapting, from it. That’s dangerous, unless there are checks. It prompted me to think about the extent to which A.I. is a mirror of society, and why we need to establish boundaries.

Potential solution:

Even the public-facing systems of AI should have strong content moderation pipelines — including realtime input filters and RLHF models for output guiding. Application developers need to test worst-case prompts and responses before deployment.

Problem 2: The Risk of Manipulation

Why I chose this issue:

Tay’s demise wasn’t a slow decline — it was a design problem, whereby users intentionally polluted its learning. What did you choose and why: I chose this because it highlights how adversaries can use AI in counterintuitve ways.

My thought process:

I initially thought that this was some prank gone bad. But it quickly became clear that it was an illustration of how the social platforms can act as hostile environments for AI. It’s as if you were trying to teach a child by putting them in the company of bullies. It emphasized the need for controlled learning environments, particularly where there is no supervision or in online learning.

Potential solution:

Experiment with Fast Iteration + Testing: Deploy, testing, and iterate in a closed-loop on simulated environments before going live. Also, incorporate adversarial robustness testing in order to imitate the hostile setup on the real environment.

So the question of moral conscience was already foregone by the time those algorithms began to operate.

Why I chose this issue:

Tay was unleashed on the Internet by Microsoft without a full ethical review of potential pitfalls. This demonstrates how deploying AI without considering human behavior can have disastrous consequences.

My thought process:

Initially, I assumed it was simply bad luck. But thinking more about it, this is more a planning problem. AI tends to not always be predictable, and when you couple that with unpredictable humans, it’s a recipe for disaster without ethical foresight. It made me reflexively think about anticipating such risk.

Potential solution:

Before launch, companies should implement ethical risk assessments like AI impact assessments (akin to privacy impact assessments). You can include subject matter experts from multiple departments to evaluate potential misuse use cases.

Critical Thinking Reflection

At first, I believed that this assignment was going to be to talk about how AI fails on the technical level — that is, bugs, or problems with the algorithm. I’ve since discovered the equally powerful role of human influence. Tay’s case taught me that AI is not born in a vacuum — it is educated by people, and if those people happen to be toxic, then the AI learns toxicity as well.

One significant assumption I reversed was thinking social media AI could self-learn uncontained. That’s clearly not the case. In the future, I will always think about not just how the AI works, but who is interacting with it and under what circumstances. Some people make jokes about how “ethical AI” is just about “coding the right model,” but it’s about predicting and preparing for real-world behavior.

Sources Used:

The Verge: “Twitter Trained Microsoft’s AI Chatbot to Be a Racist Jerk” (2016)

AI Incident Database – Case #112 – Tay AI Chatbot Incident

Wired Magazine: “The tragedy of Tay, Microsoft’s AI chatbot”

(Full disclosure: I used ChatGPT to organize case facts and brainstorm key issues. Analysis and reflection are all mine.)

6.3 Assignment: Commercial Applications of AI

Machine Learning

Indiana Wesleyan University

APRIL 2025

Introduction

Artificial intelligence (AI) was once more of a sci-fi, futuristic-type thing, than something actually came out to real life. 2025: Artificial intelligence is not only streamlining business operations but also reengineering customer experiences and information-based decision-making. This newsletter covers the applications of AI in healthcare, retail, and finance, discussing what they are currently doing and what possibilities remain.

Healthcare: Improving Patient Care and Administrative Efficiency

AI is revolutionizing progress in healthcare by enhancing diagnostic precision and automating administrative workflows.

Key Applications:

Diagnostics: Tools such as IBM Watson Health are helping by analyzing medical data to provide evidence-based treatment options for healthcare providers, reducing diagnostic errors, and customizing patient care.

Automated Billing and Claims Processing: Generative AI is being used for following up on accounts receivable and drafting structured fact-based appeals to health insurers with historical performance and insurer policy manuals.

Impact:

Currently, the usage of these AI applications is making doctors more efficient, helping healthcare providers utilising them reduce costs, and improving outcomes with faster and higher accuracy medical diagnosis and treatment plan.

Retail: Improving Customer Experience and Operational Efficiency

AI is reshaping consumer interaction and enhancing inventory management in the retail sector.

Key Applications:

Here are a few to get you started: AI-Powered Recommendation Systems AI-Powered Recommendation Systems: Such as what Amazon does to analyze customer behavior using machine learning algorithms to predict the products customers are likely to buy.

Inventory Management Optimization AI tools help predict the supply and demand of products as such, retailers can manage their stock, reducing overstock and stockouts.

Impact:

This is how AI is helping the retail industry increase sales, customer satisfaction, and operational efficiency through AI-driven personalization and efficient inventory management.

Industry: Focus on improving security and customer service

The finance industry is by far one of the most used businesses in daily life, which makes it more lucrative for AI.

Key Applications:

Fraud Detection Systems: AI algorithms are being used by financial institutions to analyze transaction data for anomalies, allowing for immediate detection of potential fraudulent activities and the reduction of financial losses.

AI-Powered Customer Service: Financial institutions are increasingly utilizing AI-driven chatbots to respond to customer queries promptly and accurately, leading to improved customer satisfaction.

Impact:

Through the use of AI applications in finance, security can be elevated, operational costs compressed, and customer service efficiency enhanced, building trust and satisfaction in clientele.

Conclusion

By 2025, the adoption of AI has become a must for business success and competitiveness. Whatever field we speak about artificial intelligence is revolutionizing it as they are scaling the personalized experience in healthcare, retail, and finance. As AI technology matures, the impact of the technology to transform business processes and customer experiences will only continue to grow, enabling the adoption of new and improved business practices.

Healthcare AI Integration - Increase in Diagnostic Accuracy

Line Chart: Quarterly AI Investment Growth

3. Pie Chart: Industry Share of AI Deployments