Human-centered researcher of emerging tech. Designer of alternative interactions. Research Fellow at The Center for Democracy and Technology. Ph.D. in HCI from Carnegie Mellon.
It has been particularly interesting and enjoyable to do research on a topic that has been close to my heart for many years #HRI #Socialagents #chatbots. This time in the context of policy at @CenDemTech. See our blog on the topic, led by Emilie Grybos:
https://cdt.org/insights/beyond-chatgpt-social-agents-the-policy-gap/
AI Accountability and the Risks of Social Interfaces
For the Tech Policy Press podcast, I spoke to:
• NTIA Senior Advisor for Algorithmic Justice Ellen P. Goodman
• @CenDemTech researcher Michal Luria (@mluria)
Listen:
https://techpolicy.press/ai-accountability-and-the-risks-of-social-interfaces/
TikTok’s center has a lot of potential. It’s very engaging by design. It is also a place where stakeholders can come together and have conversations with TikTok leaders, conversations that might not happen otherwise.
But to make the center meaningful in and of itself, especially given TikTok’s plan to open more, stakeholders should be involved early on. Learning from this launch, I hope that TikTok put resources into #codesign and #designresearch to inform future transparency centers. 8/8 🧵
Some of our findings discuss the very problems journalists have raised following their visit to the new TikTok transparency center. For example, that the information is presented in vague language, or a mismatch between the shared content and what was desired by visitors. 7/
Our prototypes explore opportunities for transparency reports to be more engaging, interactive, personalized and controllable for end-users. They are a first step towards creating more transparency, with the audience they were created for involved. 6/
https://cdt.org/insights/this-is-transparency-to-me-research-prototypes/
To create an informative center for researchers, journalists and policymakers, they all have to be involved in the process. The attractive design and the intention of transparency does not make up for the lack of stakeholder-input on what information should be shared.
At @CenDemTech, we lead design research for policy-related topics. In a recent report, we used design research to co-design prototypes of recommendation algorithm transparency reports with end-users. 5/
https://cdt.org/insights/this-is-transparency-to-me-user-insights-into-recommendation-algorithm-reporting/
Transparency is a complex concept that requires great care to design and communicate effectively. It is a perfect example of a “wicked” problem, a multi-faceted, complex problem where multiple stakeholders may have conflicting goals.
That’s where #designresearch comes in. It is a method specifically intended to address complicated and nuanced problems. Most importantly, it is human-centered, which means it includes stakeholders in various stages of the decision-making and design process. 4/
So why doesn’t the center work so well as a standalone? After all, it was created explicitly for the purpose of transparency, sharing information about recommendation algorithms (through code examples) and a demo that explains content moderation. My answer: Design Research. 3/
In their visit to TikTok’s new Transparency Center, journalists reported that few of their questions have been answered, leaving them with more concerns on the way out: https://www.theverge.com/2023/2/2/23583491/tiktok-transparency-center-tour-photos-bytedance
This was quite different from my own experience. On my visit, a group of people from civil society were invited to have conversations, ask questions, and raise concerns to a range of stakeholders from TikTok. These discussions became the key to making the visit worthwhile. 2/
Recently, TikTok launched an in-person Transparency Center in LA, where visitors can learn more about TikTok and how it operates. This is a refreshing take on the usually dense text-based reports, so why is its effectiveness limited? A post-visit reflection 🧵 1/
With all of the GPT talk about "the death of writing", can anyone agree with me that we should not care?
http://write.guyhoffman.com/why-i-dont-care-if-students-use-gpt
