The SEED Method
A methodology for growing trustworthy technology from human values
““On the one hand information wants to be expensive, because it’s so valuable. The right information in the right place just changes your life. On the other hand, information wants to be free, because the cost of getting it out is getting lower and lower all the time. So you have these two fighting against each other.” ”
PROJECT DESCRIPTIONA Path Back to Trust
Today, information on the internet, which is theoretically a free communal resource, is described by top data and media academics and policy institutions as necessary “data” to “exploit” for its value.
In the Oxford English Dictionary, ‘exploitation’ is defined as “The action or fact of taking advantage of something or someone in an unfair or unethical manner; utilization of something for one's own ends”. In describing the action of getting value from a free communal resource, those who do so have stated that such an action is unfair or unethical.
How is this possible? How did we get from ‘information’ to ‘data’? How can free information be used unethically? These questions highlight a twist in the fabric of our understanding of content on the internet; a critical shift between past ethical values and current economic and political desires brought on by technological advances, specifically generative AI training.
ABSTRACTThe SEED Method: Growing Trustworthy Technology
Using creative works for training generative machine learning models without consent represents a breach of trust in the foundational ethics of the internet. The resulting controversy surrounding the ethics of this practice reflects a fundamental mismatch in expectations between individual creators and machine learning model developers.
This discussion paper explores a path back to trust in data exchange ecosystems through ethical design and technical architectures, specifically confidential computing: federated learning within trusted execution environments (TEEs) and secure aggregation. To design this solution, I developed an interdisciplinary methodology I’m calling SEED.
The SEED methodology has four parts:
Scan— Human interviews and understanding: a literature review of relevant ethical values, expert interviews with 18 participants.
Elicit— Ethics synthesis from revealed values: Results yielded 9 key values for data exchange ecosystems and 3 critical design tensions.
Envision— Speculative design and worldbuilding using the Triptech method with 5 additional participants, resulting in defined user needs, risks and threats, and use case refinement.
Design— Strategic pathways towards implementation through regulation, economics, architectures and interfaces. The ethical values and user needs mapped to technical requirements, effective business models, user interaction design, policy interventions.
This framework and methodology can be used to restore trust in sustainable creative ecosystems, in alignment with emerging regulations such as the transparency requirements of the EU AI Act, EU Data Union Strategy, and a proposed French law presuming use of cultural artifacts for AI development. The path forward follows the 2019 European Parliament and Council’s description of the ideal data exchange ecosystem: “as open as possible, as closed as necessary”.
