Olga Scrivner

Olga Scrivner is a recognized expert in Natural Language Processing, AI fairness, and computational linguistics, with a dual Ph.D. in Computational and French Linguistics. She has delivered multiple invited talks on bias detection and mitigation in large language models (including the PD programs for K-12 educators), and co‑founded Scrivner Solutions Inc., where she leads practitioner workshops on data‑driven AI innovation. Her publications include the seminal chapter “Unveiling Unintended Systematic Biases in NLP” in Mitigating Bias in Machine Learning and her co‑authored “Safety By Design” framework (SA IEEE P3462), which will be presented at the IEEE GenAI summit in August. She has obtained Google GenAI Leader (the first cohort sponsored by Google) and Google Cloud Leader Foundational certifications. In addition, she is a part of MentorCollective, Google WTM mentorship, and REU to provide support with leadership and technical skills. As a member of both the Google Cloud Innovators and Google for Developers programs, Olga continuously deepens her expertise in Gemini, Agents Framework, and next‑generation generative AI tools, bringing hands‑on, up‑to‑the‑minute training directly into her skill-building workshops. As a Consultant at OpenAI, Olga is continually exposed to the frontiers of generative AI,

An active IEEE Women in Engineering (WIE) Executive Committee member for Central Indiana, Google Women TechMakers (WTM) Ambassador, a member on board of directors for American Council on Education Women Indiana Network (ACE WIN), and former faculty fellow for CEWIT (Center of Excellence for Women in Technology) and learning analytics fellow at Indiana University Olga has developed multiple hands‑on seminars in ethical AI, programming workshops, and immersive data‑visualization techniques. She has also organized multiple virtual panels on Leadership (with the IEEE president-elect), cybersecurity (with local businesses in Indianapolis, including Cummings and Rolls-Royce), both virtual and in‑person.

Contact: obscrivn@iu.edu

Website: https://scrivner-solutions.com/

Abstracts

A Safety-by-Design Lifecycle for Generative AI

As generative AI proliferates, models can inadvertently generate or facilitate access to child sexual abuse material (CSAM). Yet, developers face a complex, evolving patchwork of child-protection regulations, spanning the US Children’s Online Privacy Protection Act (COPPA), the EU AI Act and Digital Services Act, and the UK Online Safety Bill, which complicates the design and implementation of safety pipelines and introduces compliance gaps. We present a unified, region-agnostic “safety-by-design” lifecycle framework that embeds child-safety guardrails natively at every stage of the ML pipeline. From data curation to model training, deployment, continuous monitoring, and eventual system retirement, our approach makes compliance an intrinsic part of the process, rather than a retrofitted addition. Our safety-by-design lifecycle framework (IEEE P3462) (1) consolidates global child-safety requirements into a single reference model, (2) delivers a clear, step-by-step playbook for immediate operationalization, and (3) offers strategies for evolving alongside both regulatory updates and advances in generative model architectures. Crucially, this framework ensures that child protection is not an afterthought but an integral part of every phase within the lifecycle of generative AI systems.

Keywords: Safety-by-Design Framework, CSAM, generative AI child-protection standards

Balancing Innovation and Accountability: Can AI be Regulated?

The rapid advancement and diffusion of artificial intelligence (AI) technologies are transforming many industries, from Human Resources to Finance and Supply Chain. Generative AI's capabilities, particularly in imaging, speech, and NLP, are accelerating the adoption of AI-powered applications, leading to the projected 25% (300B) increase in global spending.

There are also concerns about AI technologies' potential risks and ethical implications. The lack of comprehensive and standardized AI regulations leads to varying levels of oversight in different countries. Many collaborative initiatives from government, policymakers, industry leaders, and researchers have been launched to address the concerns regarding data privacy, algorithmic biases, transparency, and the impact of AI on labor markets.

This presentation will review several initiatives (AIAAIC, FutureWorks) and compare AI regulations in the US and EU, highlighting distinct approaches and their potential impacts on AI adoption.

Building Ethical, Scalable Generative AI: From Bias Auditing to Cloud‑Native Deployment

As generative AI and agent‑based architectures reshape computing, organizations must balance rapid innovation with robust safety, fairness, and governance. This workshop will guide attendees through a practical framework for deploying models at cloud scale using platforms such as Google Cloud AI Studio and Vertex AI. Attendees will learn hands‑on techniques for prompt engineering, learn about vibe-coding, MLOps best practices, and workflows.