Description
- This is our response to “it’s not my domain” syndrome
- Demanding accountability to the corporate almighties
- Care for the human connection, figuring the value constellation.
Further information
Accountability means being answerable for one’s choices, actions and expectations. It also applies to how answerable services and systems are and how these should ‘account for’ their affordances in intelligible ways.
- Who or what is accountable?
- AI is redefining concepts that were long thought resolved. Just how can an algorithm be accountable, and to what?
- Does a program made by humans have accountability for its decisions?
- Can the link between AI decision and human reason be maintained as technology becomes more autonomous?
- Who is accountable for the autonomous drone that shoots the wrong person?
- How are we accountable when the stakes of our lives shift into automated systems?
To start addressing this shifting landscape, we begin with the narrative that AI is, and must be, entwined with trust and care. We accept the necessity of our being vulnerable, of opening possibilities that we cannot fully predict. At the same time, we demand that those links we make to AI, be worthy of trust, that we care for them and that they care for us. In the midst of massive corporate control of AI innovation, the task ahead is to bring this back down to earth. To do so with care, according to de la Bellacasa, is to engage in ‘a manifold range of doings needed to create, hold together, and sustain life and continue its diverseness.’ (de la Bellacasa, 2017, p. 70). Included in this diverseness is a recognition that AI does not necessarily care in the same way we do. Holding AI accountable means opening new domains, and new ways of doing care.
Resources
- Kate Crawford and Trevor Paglen, 2019 ‘Excavating AI: The Politics of Images in Machine Learning Training Sets’ www.excavating.ai.
- No Justice, No Robots: An Open Letter From Robotics Researchers. Accessible at: https://nojusticenorobots.github.io