Abstract: Decent general use models have used generally obtainable data to prove the utility of many compelling use cases for AI, such as speech recognition, relevance and image recognition. Now that the utility of these capabilities has been proven and the demand to solve more specific business problems unlocks additional investment dollars for AI, I will make the case that these use cases will increasingly focus on refining models with less accessible and more specific data. From rare events to less well-represented demographics, the number of edge cases will increase. Meanwhile increasing regulation, increased privacy protection expectations, tighter infosec requirements and ultimately model explain-ability will challenge us to systematically tighten governance and data management, as well as up the ante for higher fidelity in synthetic data.
Bio: Erik passionately advocates for tomorrow's solutions, with a keen focus on pragmatically getting there today. With over 20 years' experience in operations, sales, and engineering in the language services and data annotation industries, Appen's VP of Enterprise Solutions brings a wholistic approach to building creative fit-for-use solutions from discovery through delivery. Erik's broad background in business strategy and people-centric leadership is focused on building more compelling and ethical value propositions for clients, people, and shareholders. Erik has an MBA, an MS in Management and Leadership, and an BA in Psychology.
