The generative AI industry is a globally relevant target for Open Terms Archive, considering on the one hand that it is so emergent it is mostly unregulated and will thus evolve very fast; and on the other hand that its potential for societal transformation, disruption and harm are as immense as the new opportunities it will offer.
At the same time, several national and global regulatory proposals on AI exist, but none have been adopted yet: UNESCO’s “Ethics of AI”; Council of Europe’s “Convention of AI”; EU’s “AI Act”; White House’s “Blueprint for an AI Bill of Rights”; China’s “AI rulebook”; OECD “AI Principles”… Legal challenges on personal data handling arise, starting with Italy and France. The impact of these upcoming constraints will be observable in the evolution of the terms, and legislative drafting should be informed by the latest state of business affairs.
The generative AI theme also seems a perfect fit for trialling an expansion of targeting to involve individual contributors. First of all, it is an extremely visible topic that both fascinates and terrifies many people, making it easy to gain attention and involve individuals, as opposed to data privacy rights, for example. Second, it is extremely rapidly evolving, which means that it is very hard for any centralised actor to be aware of all emerging services, and it is thus very relevant to obtain that knowledge from the crowd.
We will thus run two cycles of crowdsourcing on slightly different themes, building on the learnings from the first one to optimise the second and derive generic knowledge to deploy this approach more systematically to other Open Terms Archive targets. The first one will be specifically focused on the generative AI services such as OpenAI, New Bing, Dall-E or Stable Diffusion. The second one will focus on the content sharing platforms that are impacted by a massive influx of synthetic contributions and try to adapt their policies, such as StackExchange, Artgram or Shutterstock.
The resulting public archives will increase scrutiny and serve as a historical record of how the industry evolved and how it progressively transformed society. This database of terms will provide a basis for the design of targeted regulation, enabling lawmakers to understand the impact of each change over time, to measure the effectiveness of their efforts in changing policies, and to assess compliance in a much faster way. It will empower the whole society to identify specific changes, react to them and lobby for alternatives, as described in our impact model.
The Open Terms Archive core team has a full project plan to develop the tools and processes necessary to support the crowdsourcing and tracking of documents, including detailed budgets and timelines. We estimate the full costs to be around 200k€ and are looking for supporters to reach that amount, either through funding or through in-kind contributions. If you can participate or know of relevant calls, please contact us!
If you are an individual or any other type of entity and willing to participate in the crowdsourcing when it opens, please drop us a line describing your potential availability so we can register your interest and inform you of the opening.
Open Terms Archive publicly records the terms of services of online services in different languages and countries several times a day, increasing their readability and highlighting their changes. It is designed as a tool to steer platform governance towards increased accountability, improved legislation and stronger regulatory compliance. Read more on our impact model.