Fed up with Big Tech, communities turn to data collectives for control

Communities are increasingly turning to data collectives and cooperatives as a means of regaining control over the collection, management and distribution of their data, in response to the dominance of a few big tech companies in the generative artificial intelligence industry. These non-profit organizations provide a platform for communities and individuals to control their unique data sets, which may otherwise be ignored or misused by large corporations.

Companies like Meta, OpenAI, Google and Anthropic have scraped vast amounts of data from the internet to train their AI models, often without consent and accused of using copyrighted material. Some countries are attempting to balance the need for good data with compensation for creators, while data collectives offer a blueprint for responsible use of data, ensuring that the value goes to those generating it.

Data collectives cover a wide range of applications, from enabling farmers in India to trace and market their produce, to helping fishers in Latin America manage and benefit from their catch records. They also provide a solution for low-resource language communities, where data sets can be used to train small AI models and create speech recognition tools.

For example, Meesum Alam, a researcher from Pakistan, documented endangered languages and contributed voice data sets to Mozilla Data Collective, where they have been used by companies like Meta to develop speech recognition tools. The communities set strict terms of use for the data sets, restricting their use to research and non-commercial purposes.

Other frameworks, such as data trusts, data unions, data commons and data donation schemes, are also being used to manage data more responsibly. In Africa, the Nwulite Obodo Open Data License enables creators, communities and researchers to share data sets without relinquishing their right to benefit from them.