It’s been tough to stay focused, with everything that’s happening in the world right now. But these developments have sparked some reflections around power and information in me, including how they relate to my regional Brazilian reality and my work around data.
IGF
I attended a few sessions of the 2023 Internet Governance Forum last week. I haven’t been able to attend IGF in person (like I have in its regional and national editions), but I understand it as an essential forum for conversations around digital rights and data and appreciate its hybrid format.
During the town hall “Dare to Share: Rebuilding Trust Through Data Stewardship”, an argument for a stronger emphasis on enforcement capacity in data governance was made. The Brazilian Data Protection Authority is an interesting case in this regard, as it has adopted an approach to compliance based on education and communication, through opinions and concept notes, rather than sanctions. According to the institution, sharing knowledge has proven more beneficial in guiding stakeholders through compliance with the Brazilian Data Protection Law.
On the opposite side of the spectrum of examples, the inadequate enforcement practices in the US were mentioned as an example of the distinction between having legislation in place and effectively implementing and enforcing it. This fundamental question of the sociology of law (“ought” vs. “is”) is commonly acknowledged in Brasil: whenever a law in Brasil is not, or cannot, be meaningfully enforced, we say that it “did not take hold” (a lei não pegou).
The particular case of the US also poses important questions about power. And it reminded me of my first international experience as a researcher. During a conference in the country for early career researchers, an Ivy-League American PhD candidate presented her work, which consisted of analysing the US contributions during the public consultation to the draft bill that would later become GDPR. Her main conclusion: the country’s contributions had little impact in shaping the regulation, a result similar to the one of “David and Goliath.” Associating Europe and the (many) countries all over the world that adopted GDPR-inspired regulations with a smaller, weaker character reveals (amongst other things) a lot about how power dynamics have impacted tech and challenged the distinction between the digital and the real world.
This analytical lens (i. e. power) also made me challenge the moderator’s analogy for data sharing, according to which data sharing should be viewed as a collaborative effort, similar to a dinner with friends, where everyone brings a dish to share.
Overall, it was interesting to see how conversations around enforcement can also make the case for participatory and deliberative data governance practices: they can facilitate implementation and stakeholder buy-in.
A call for collective and inclusive decision-making and assessment was also made during the session “Evolving AI, evolving governance: from principles to action”. And it was furthered in the session “Decolonise Digital Rights: For a Globally Inclusive Future” with a specific focus on language and how AI current systems are undermining the preservation of unique cultures, traditions, and identities associated with lesser-known languages (incidentally, language differences is something Emily and I have been discussing a lot).
It was especially encouraging to see this argument in the latter panel after one of the speakers argued for individual (and insufficient, I might add) measures to ensure data subjects’ security - such as limiting one’s social media presence and critical thinking before agreeing to privacy terms and conditions that may require sharing personal data. A permission-based approach such as this seems to overlook power asymmetries and emphasise the (exclusionary) data and technical aspects of data flows rather than citizens’ rights.
Generative AI for not dummies
Outside CONNECTED BY DATA, I was invited to give a lecture on the possible uses of generative AI for empirical legal research. I always appreciate opportunities to talk about technology and data with people from outside the field. So, I accepted the invitation but proposed an approach that was different in two main ways.
First, I decided to put together and facilitate a workshop, rather than give a formal, traditional lecture. Getting to know different designs for sessions, conferences, and convenings has been one of the most rewarding learnings during my time at CONNECTED BY DATA. Since, generally speaking, people from the Brazilian legal field seem to be falling behind on these alternative, participatory, interactive methods, I saw the experience as an opportunity to both share and explore some of the things I’ve been seeing.
Secondly, as these technologies are new and continue to change, it was important for me to focus on the tools and information that would allow participants to assess for themselves the possible uses of AI systems, in the long haul. All participants had heard of ChatGPT, but none of them knew how it works, which is a prerequisite for using it properly and effectively - one example of misuse due to lack of understanding are the many cases of teachers and professors using ChatGPT to spot plagiarism in Brasil.
After understanding which data ChatGPT uses and how it uses it, participants were able to raise themselves some concerns and point out some limitations, from inaccurate information to the above-mentioned issues around language and stereotypes - I used a ChatGPT-generated comedy sketch about Brasil for this last one.
The final activity was for them to try using ChatGPT to create a solid empirical research project: in some groups, they did so by choosing a topic that related to the research topics of all members, while in others they chose the topic of one of the members to later compare the result with the research and findings of that member.
My main learning from this experience relates to Helena’s discomfort with the term “literacy”, and her point on how critical thinking and access to information are the elements needed to empower people to make decisions about tech.