A handful of companies are using insurmountable amounts of resources (from land, to water, labour and money, to name a few) to build tools that frame one particular vision of what “AI” should and could be. Their framing espouses a technodeterminist future, in which the development of large, resource-intensive, general-purpose AI models is inevitable. In this session, we aim to highlight how those of us sitting outside of Big Tech have a choice in what it can look like and whether (and how) it can serve us. Together, we will explore pathways to moving from passive observers of AI to actively shaping it and reclaiming control, in opposition to technodeterministic views.

We will demonstrate how purpose-driven AI tools, specialized for specific tasks, can be viable alternatives to large, all-purpose models; we will discuss how local, small AI can mean less dependence on cloud computing and data centers, and the potential of edge computing on small devices. We will also discuss small ways we can be intentional about our day-to-day use of AI, exercise our agency in refusing AI-creep into our lives, and push back harder against actors who defend the indiscriminate adoption of certain AI models. Our session will offer multiple modes of sharing and reflection and will use a series of prompts during panel discussion to help participants reflect as the discussion unfolds. Additionally speakers will bring their own key questions to help scaffold and situate participant learning to help attendees think about the change they wish to compel.

Host institution: Open Environmental Data Project (OEDP)

Speakers:

• Cathy Richards, OEDP
• Quito Tsui, The MERL Tech Initiative
• Olivia Johnson, The Engine Room

Facilitator: Bárbara Paes