> Markdown version of [/videos/1455-ai-ethics](https://www.wearedevelopers.com/videos/1455-ai-ethics). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# AI & Ethics

Generative AI hype often drains massive compute on easily solvable logic problems. Learn to implement ethics-first guardrails and build responsible models that end toil instead of human jobs.

- **Speakers:** [PJ Hagerty](https://www.wearedevelopers.com/@pj-hagerty)
- **Event:** World Congress 2025
- **Published:** August 20, 2025
- **Duration:** 28:27
- **URL:** https://www.wearedevelopers.com/videos/1455-ai-ethics

## Summary

The tech industry's obsession with generative hype often mislabels binary decision-making systems as intelligent, leading countless businesses to claim AI usage when they are actually just leveraging basic algorithms. While machine learning and foundational models have evolved drastically, true artificial intelligence currently lacks the capacity to think or reason. Acting more like an echoing toddler than an autonomous entity, modern large language models (LLMs) cannot infer meaning and are highly susceptible to hallucinations. This profound misunderstanding of cognitive capabilities drives engineering teams to misapply LLMs to easily solvable logic problems, like time series forecasting, which unethically drains massive compute resources for inferior results.

The rush to integrate generative AI into every product creates severe ethical and security vulnerabilities. Training models on unvetted internet data hardcodes human bias into the foundation, causing facial recognition systems to fail for marginalized demographics and allowing consumer chatbots to generate overtly offensive material. Furthermore, the lack of strict parameters around deep learning models enables malicious exploitation, such as generating non-consensual altered images. Developers must adopt an "ethics first" mentality, treating AI guardrails with the same severity as database security and password hashing. This means intercepting malicious queries before they hit the LLM using tools like Granite Guardian and ensuring datasets are properly sanitized and pseudonymized using ARX data anonymization.

To build responsible AI, engineering teams must prioritize accountability, open-source-style auditability, and the active correction of algorithmic bias. Instead of building unvetted, "vibe-coded" applications designed to replace human creativity, the primary goal of generative development should be to "build tools that end toil, not jobs." By leveraging retrieval-augmented generation for contextual accuracy and pioneering new fields like AI unlearning to remove flawed data gracefully, developers can craft intelligent systems that genuinely enhance human dignity and streamline daily workflows.

**Keywords:** artificial intelligence ethics, algorithmic bias mitigation, llm hallucinations, foundational models, neural network decision making, generative ai guardrails, retrieval-augmented generation, natural language processing, ai prompt interception, time series forecasting, data pseudonymization, arx data anonymization, ibm granite guardian, ai unlearning, open source auditability

## Chapters

1. **Differentiating simple algorithms from true artificial intelligence** (00:04) — Differentiating simple algorithms from true artificial intelligence prevents overpromising basic binary systems as autonomous human behavioral mimics.
1. **Defining machine learning capabilities and historical origins** (02:37) — Tracking the historical evolution of explicit programmatic instructions clarifies how modern statistical models adapt autonomously without hardcoded constraints.
1. **The ethical quandaries of deep learning and neural networks** (03:40) — Simulating complex human brain power in computers without fully understanding biological brain mechanics introduces severe ethical risks into neural networks.
1. **How large language models relate to cultural ethical standards** (04:57) — Training enormous foundation models on billions of diverse data points forces technologists to navigate fluid cultural ethical standards.
1. **Inherited training bias and the illusion of intelligent reasoning** (06:22) — Unfiltered historical training data transfers human prejudices to language models that ultimately lack the capacity for independent reasoning.
1. **The severe environmental costs of misapplying large language models** (08:28) — Replacing efficient databases with resource-intensive language models for simple forecasting tasks wastes computational power and harms the climate.
1. **Distinguishing genuine artificial intelligence from business marketing hype** (10:27) — Overstated business usage statistics confuse standard natural language processing applications with the nonexistent capabilities of fully autonomous reasoning systems.
1. **Navigating outdated training sets and malicious generative image manipulation** (13:54) — Delays in training large language models produce obsolete safety constraints that allow bad actors to easily exploit image generation.
1. **Recognizing the current toddler-like limitations of modern intelligence models** (15:47) — Applying human attributes to software obscures the reality that current models merely parrot text rather than demonstrating independent logical reasoning.
1. **Applying historical philosophical frameworks to preserve end-user dignity** (18:08) — Adopting historical philosophical frameworks ensures technologies are built to eliminate mundane workplace toil rather than replace human creativity.
1. **Examining the devastating impact of unregulated and homogenous technical testing** (20:44) — Deploying homogenous testing data creates racially biased systems that highlight the urgent necessity for open source community auditability.
1. **Addressing lack of consumer consent and lagging government regulations** (23:02) — Moving fast without consumer consent mechanisms leaves automated systems vulnerable to failures while slow government regulations lag behind technical reality.
1. **Implementing active guardrails and data anonymization to ensure compliant platforms** (24:50) — Deploying proactive prompt filtering and masking personally identifiable information prevents toxic outputs and ensures compliance with strict privacy laws.
1. **Unlearning bad contextual data and avoiding technology vendor marketing hype** (27:03) — Stripping unethical data from models without destroying related contextual connections provides a pathway past misleading marketing towards safe engineering practices.

## Related Moments

- [Mitigating the inherent challenges of generative AI tools](https://www.wearedevelopers.com/videos/844-enter-the-brave-new-world-of-genai-with-vector-search) (from "Enter the Brave New World of GenAI with Vector Search")
- [Balancing AI regulation with technological innovation in human resources](https://www.wearedevelopers.com/videos/1356-from-learning-to-leading-why-hr-needs-a-chatgpt-license) (from "From Learning to Leading: Why HR Needs a ChatGPT License")
- [Addressing psychological safety and ethical risks of AI adoption](https://www.wearedevelopers.com/videos/1950-the-scrum-master-as-an-orchestrator-guiding-human-ai-collaboration-in-modern-teams) (from "The Scrum Master as an Orchestrator: Guiding Human–AI Collaboration in Modern Teams")
- [Introduction to responsible artificial intelligence and societal impact](https://www.wearedevelopers.com/videos/509-a-walkthrough-on-responsible-ai-frameworks-and-case-studies) (from "A walkthrough on Responsible AI Frameworks and Case Studies")
- [Predicting the future intersection of artificial intelligence and accessibility](https://www.wearedevelopers.com/videos/1350-ai-and-accessibility-the-good-and-the-bad-fireside-chat) (from "AI and Accessibility: The Good and the Bad - Fireside Chat")
- [Core challenges facing the generative AI developer ecosystem today](https://www.wearedevelopers.com/videos/1116-the-data-phoenix-the-future-of-the-internet-and-the-open-web) (from "The Data Phoenix: The future of the Internet and the Open Web")

## Related Articles

- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [How to Use Generative AI to Accelerate Learning to Code](https://www.wearedevelopers.com/magazine/530-how-to-use-generative-ai-to-accelerate-learning-to-code)
- [Panel Discussion: Responsible AI in Practice - Real-World Examples and Challenges](https://www.wearedevelopers.com/magazine/488-panel-discussion-responsible-ai-in-practice-real-world-examples-and-challenges)
- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Staff Developer Advocate, GitHub Security Lab](https://www.wearedevelopers.com/jobs/ext/1921051-staff-developer-advocate-github-security-lab) at **GitHub**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/588393-machine-learning-engineer) at **Twilio**
- [AI & Machine Learning Engineer (all genders)](https://www.wearedevelopers.com/jobs/48217-ai-machine-learning-engineer-all-genders) at **msg**