Reports & Publications


Delegated Regulation on data access provided for in the Digital Services Act

In order for the Digital Services Act (DSA) to achieve its goal to facilitate research that identifies and addresses new systemic risks on very large online platforms and very large online search engines, the following actions should be taken: 1. An independent, intermediary body should be established to assist the DSCs in vetting researchers and research projects. The DSCs may lack the expertise and capacity to adequately review researchers and research projects in a timely manner. Thus, the independent body can provide robust peer review while streamlining and quickening the review process for the DSCs and assist in ensuring compliance with relevant laws and regulations (e.g., GDPR). The European Digital Media Observatory (EDMO) has launched a working group to establish such a body. 2. Establish requirements for high-value API access.

California Executive Order on Generative AI

Generative AI is a new technology that is not well understood, even by experts. It can perform surprisingly well in image and text generation, but it is prone to bias and errors. It is VERY important that those considering its use learn about the many limitations of generative AI and keep these in mind when making decisions about its use.
California should develop resources for state employees considering use of generative AI. The government must support the development of training materials for government personnel in generative AI tool capabilities, how to use them appropriately, and how to evaluate their effectiveness, including mechanisms for personnel to report when
The state’s efforts and the executive order should not only focus on generative AI, but on a broad range of AI methodologies, such as machine learning, reinforcement learning, deep learning and more. The state is currently using many of these methodologies and would benefit greatly from the execution of a similar process as outlined in the Governor’s executive order on generative AI.

NIST AI Risk-Management Standards Profile for General-Purpose AI Systems (GPAIS) and Foundation Models

Increasingly multi-purpose AI systems, such as cutting-edge large language models or other “general-purpose AI” systems (GPAI or GPAIS), “foundation models,” generative AI, and “frontier models” (typically all referred to hereafter with the umbrella term GPAIS except where greater specificity is needed), can provide many beneficial capabilities but also risks of adverse events with profound consequences. This document provides risk-management practices or controls for identifying, analyzing, and mitigating risks of GPAIS. We intend this document primarily for developers of large-scale, state-of-the-art GPAIS; others that can benefit from this guidance include downstream developers of end-use applications that build on a GPAIS platform. This document facilitates conformity with or use of leading AI risk management-related standards, adapting and building on the generic voluntary guidance in the NIST AI Risk Management Framework and ISO/IEC 23894, with a focus on the unique issues faced by developers of GPAIS.

Amicus Brief in Supreme Court Case Gonzalez v. Google

CITRIS Policy Lab founding Director Brandie Nonnecke is part of a group of expert technologists that filed an amicus brief urging the U.S. Supreme Court to have Section 230’s liability shield apply to claims against interactive computer service providers based on their recommended content.

The brief demonstrates that if recommendations are not shielded by Section 230, providers will be discouraged from using novel algorithms to help users find content. It also explains that such a holding would create strong incentives for providers to limit speech for fear of potential liability.


Harass, Mislead, & Polarize: An Analysis of Twitter Political Bots’ Tactics in Targeting the Immigration Debate Before the 2018 U.S. Midterm Election


This study investigates the interaction and messaging tactics of political Twitter bots before an election. We analyzed the strategies of influential bots seeking to affect the immigration debate before the 2018 U.S. midterm elections. Our findings reveal that the 10 most influential bots in our dataset all presented an anti-immigration viewpoint, and both posted original tweets and retweeted other bot accounts’ tweets to give a false sense of authenticity and anti-immigration consensus. Bots’ messages relied heavily on negative emotional appeals by spreading harassing language and disinformation likely intended to evoke fear toward immigrants. Such accounts also employed polarizing language to entrench political group identity and provoke partisanship. Our findings help to understand the interaction and messaging tactics employed by political bots and suggest potential strategies that may be employed to counter their effectiveness.

Women’s Reproductive Rights: Computational Propaganda in the United States


Women’s reproductive rights have been a divisive political issue in the United States for decades. While social media holds great potential to support bipartisan engagement around politically contentious issues like reproductive rights, these platforms can also be co-opted by nefarious actors to perpetuate polarization. Using the Twitter Search and Stream APIs, we captured over 1.7 million tweets corresponding to 463,261 unique handles. Using a combination of social network analysis, bot detection, and qualitative coding this research explores the role of bots in spreading disinformation, harassment, and divisiveness on women’s reproductive rights.


Blockchain, Digital Identity, and Health Records: Considerations for Vulnerable Populations in California


What opportunities might blockchain-based digital ID and health records management systems offer to improve services for housing-insecure and homeless populations in California? And what risks unique to this cohort should be considered before launching such programs?

Through a combination of expert interviews, surveys with state chief information officers, and findings from a year-long process with the State of California’s Blockchain Working Group, we find that blockchain-based identity and health records management systems may indeed offer advantages for unhoused populations in California. However, issues with user authentication, cost, and trust must be addressed for these benefits to be realized.

Responsible Digital ID: Effects of Data Governance Policies and Practices on Human Rights


An estimated 1 billion people lack formal identification globally, restricting their ability to meaningfully participate in the economy and society. In response, national digital identity systems are rapidly being deployed by a variety of actors and institutions to provide individuals with formal means of establishing their identity. While these systems have the potential to hold great value to individuals, lack of sound data governance policies and practices present risks to individual civil and political rights. Through an analysis of national DID systems in Argentina, Estonia, Kenya, and China, we investigate how data governance policies and practices affect civil and political rights within the areas of data protection, political participation, and inclusion of diverse ethnic identities. We conclude with priority recommendations for national digital ID system data governance policies and practices that should be implemented to support civil and political rights.

Facing the Future: Protecting Human Rights in Policy Strategies for Facial Recognition Technology in Law Enforcement


Facial recognition technology (FRT) is gaining traction in law enforcement as a tool to identify persons of interest in criminal investigations. However, FRT leverages a uniquely sensitive biometric trait that is both immutable and always exposed to the public, which means that unregulated use of FRT in law enforcement creates risk for human rights. This research serves as a resource for discourse and policymaking around FRT by providing a systematic three-dimensional policy analysis framework to assess to which degree regulatory policies safeguard the most relevant human rights in the context of FRT: privacy, equity or non-discrimination, and due process. The analysis draws on qualitative methods, including a literature review, expert interviews, and archival research to operationalize each concept in measurable sub-variables and apply the framework to two case studies of mature democracies active in FRT use and committed to protecting civil liberties, the UK and the US.


Building on UC Broadband: Strategies to Improve Broadband Access Throughout California 


Access to high-speed broadband has become an essential enabling technology for healthy, productive lives. Because of its extensive geographic footprint, vast infrastructure, technical capacity, and deep ties to surrounding communities, the University of California (UC) can serve an important role in supporting broadband infrastructure development and adoption initiatives. In doing so, UC can leverage its unique capabilities to help the state of California close the digital divide, while simultaneously supporting its core mission to provide world-renowned teaching, path-breaking research, and meaningful public service. This report offers four guiding strategies and corresponding recommendations UC should consider. 


Technologies of Pandemic Control: Privacy and Ethics for COVID-19 Surveillance


While technology could assist in mitigating the spread of COVID-19, it may do so at the risk of eroding significant data privacy protections and civil liberties. Through an analysis of four technologies aimed at mitigating the spread of COVID-19 in the U.S., we investigate how these technologies work, the privacy and ethical questions they raise, the legislation governing their use, and their potential consequences now and in a post-pandemic future.

Health, Wildfires & Climate Change in California: Recommendations for Action


In Spring 2019, the Center for Information Technology Research in the Interest of Society (CITRIS) and the Banatao Institute collaborated with the California Institute for Energy and Environment to bring together UC Berkeley faculty and students with practitioners from local governments, state agencies and community organizations to examine the health impacts of wildfires and to identify solutions to better protect health in future events.


Artificial Intelligence for Economic Development in Nigeria


Artificial Intelligence (AI) has been identified as a potentially less costly and far-reaching tool in the work to eradicate poverty and achieve the UN Sustainable Development Goals. This paper assesses various sectors of the Nigerian economy and highlights technological innovations that can address some of the development challenges faced by the country. The paper concludes with an ethical framework to guide the development and deployment of AI technologies in a sustainable and inclusive manner.

Responsible Artificial Intelligence: Recommendations to Guide the University of California’s Artificial Intelligence Strategy


Launched in August 2020, the interdisciplinary “UC Presidential Working Group on Artificial Intelligence” was composed of 32 faculty and staff from all 10 UC campuses. The Working Group met monthly from August 2020 to August 2021 and developed a final report containing a set of “UC Responsible AI Principles” and recommendations to guide the responsible procurement, development, implementation, and monitoring of AI as may be implemented in UC operations.

The Future of Public Sector Work: Human-Centered Technology and Policy Strategies 


The Future of Public Sector Work explores the application of emerging technologies, especially AI, in three public sectors: K-12 education, social services, and law enforcement. The report investigates not only the effects of these technologies on efficiency and effectiveness of work, but also on equity for those served. The report concludes with priority technology and policy strategies to maximize benefits.

Putting AI to Work: Technology and Policy for Enabling the Workforce of the Future


Putting AI to Work explores ways AI can be designed and deployed to enhance and augment human labor, especially for aging populations and individuals with disabilities.

Inclusive AI: Technology and Policy for a Diverse Urban Future


Inclusive AI explores the benefits and risks of AI-enabled technologies on social, political, and economic inclusion in the urban context and provides policy recommendations for the private and public sectors.


When the Streetlights Come On: How “Smart Cities” Are Becoming a Surveillance State


Implementation of “smart city” technologies are gaining increasing popularity as cities throughout the country seek to improve efficiency and livability. Yet cities are often not transparent about the risks these technologies pose to residents.t his report provides concrete recommendations grounded in critical race, feminist, and disability rights scholarships that emphasizes the importance of centering marginalized voices within smart cities.

IoT & Sustainability: Practice, Policy and Promise


IoT & Sustainability explores applications of IoT for sustainability in the energy, water, and transportation sectors and offers recommendations to city-level officials seeking to implement IoT technologies within these domains.



Nonnecke, B., Mohanty, S., Lee, A., Lee, J., Beckman, S., Mi., J., Panpairoj, T., Ancheta, J., Martinez, H., Oco, N., Roxas, R., Crittenden, C., & Goldberg, K. (2018, October). Malasakit 2.0: A participatory online platform with feature phone integration and voice recognition for crowdsourcing disaster risk reduction strategies in the Philippines. Proceedings of the IEEE Global Humanitarian Tech Conference (GHTC). San Jose, CA.



We present Malasakit 2.0 (meaning “sincere care” in Filipino), an inclusive, multilingual participatory online platform with feature phone integration for collecting and analyzing quantitative and qualitative textual and audio data on disaster risk reduction (DRR) strategies. Malasakit 2.0 introduces interactive voice response (IVR) to support collection of audio data via feature phone. Malasakit utilizes peer-to-peer collaborative evaluation to identify and prioritize local DRR strategies. We present results from four field tests where 261 participants provided 1,582 evaluations of current DRR strategies, and over 950 peer-to-peer evaluations on 280 textual and audio suggestions for how local government (i.e., barangays) could better support vulnerable groups (e.g., elderly, women, children, and people with disabilities) during typhoons and floods. Results suggest that individuals who engage in disaster drills are also likely to participate in their barangay’s clean-up drives to reduce flooding risk by clearing drainage pathways and that those who participate in disaster drills are also likely to have enough emergency supplies for a disaster. High-rated suggestions for DRR strategies for vulnerable groups emphasize the need for communities to establish response teams that prioritize reaching out to vulnerable groups for coordination during a disaster. Malasakit can be accessed at

Nonnecke, B. Mohanty, S., Lee, A., Lee, J., Beckman, S., Mi., J., Krishnan, S., Roxas, R., Oco, N., Crittenden, C., & Goldberg, K. (2017, October). Malasakit 1.0: A participatory online platform for crowdsourcing disaster risk reduction strategies in the Philippines. Proceedings of the IEEE Global Humanitarian Tech Conference (GHTC). San Jose, CA.



We describe Malasakit 1.0 (meaning “sincere care” in Filipino), a customizable participatory assessment platform that collects and streamlines quantitative and qualitative analyses and insights of disaster risk reduction (DRR) strategies. While supervised classification approaches offer opportunity to understand qualitative textual suggestions, those methods break down in areas like the Philippines, home to hundreds of dialects and regional language nuances in varying socioeconomic contexts. Instead, Malasakit uses dimensionality reduction and peer-to-peer evaluation on qualitative textual suggestions to identify locally appropriate DRR strategies. We present results from 12 field tests conducted in eight distinct geographic locations in the Philippines. 998 participants provided 7,249 evaluations on flood and typhoon preparedness and 2,675 peer-to-peer ratings on 907 textual suggestions for how local government could improve DRR strategies. Results suggest that female participants are more confident than males in their community’s ability to recover from a major typhoon. High-rated textual suggestions focus on issuing immediate early warnings and cleaning drainages to reduce flooding. Malasakit can be accessed at

Meneses, M., Nonnecke, B., Martin del Campo, A., Krishnan, S., Patel, J., Kim, M., Crittenden, C., & Goldberg, K. (2017) Overcoming citizen mistrust and enhancing democratic practices: Results from the e-participation platform ‘Mexico Participa.’ Information Technologies and International Development (ITID), 13, 138- 154.



This article reviews the experiences, obstacles, and lessons learned from development and deployment of the México Participa e-participation platform as a case study for future platforms, both in Mexico and in transitional democracies with similar sociopolitical characteristics such as pervasive distrust of public institutions and limited civic participation. México Participa was released three months before the June 2015 midterm Mexican presidential election. Although the platform continues to operate, this article focuses on the period leading up to the election. 3,054 participants offered 336 suggestions and provided 14,033 peer-to-peer assessments. A postelection survey highlighted the need for a platform such as México Participa to be continually available to sustain citizen evaluation of government performance and to promote transparency and accountability.

Epstein, D. & Nonnecke, B. (2016). Multistakeholderism in praxis: The case of the regional and national IGF initiatives. Policy & Internet, 8(2), 148-173. doi: 10.1016/j.telpol.2015.12.005



The growing phenomena of regional and national Internet Governance Forum (IGF) initiatives offer an opportunity to look into how various interpretations of the multistakeholder model play out in different cultural, political, and economic settings. The variety of ways in which the multistakeholderism is enacted are expressed through the organizational structures and procedures of these events, their funding mechanisms, their agendas and formats, the kind of participation they attract and enable, and their potential influence on the national, regional, or global Internet governance debates. This article is a systematic attempt to map out regional and national IGF initiatives with an emphasis on how the multistakeholder model is playing out in various contexts. This analysis builds on existing dispersed documentation of these initiatives, transcripts from meetings (such as global IGF interregional dialogs), and interviews with individuals engaged in facilitation of regional and national IGF initiatives. The goal of this exercise is to offer an empirically grounded framework for thinking about the emerging models of multistakeholder governance.

Nonnecke, B. (2016). The transformative effects of multistakeholderism in Internet governance: A case study of the East Africa Internet Governance Forum. Telecommunications Policy, 40(4), 343-352.



Internet governance issues tend to be contentious, multifaceted, and interconnected among various stakeholders. Since Internet governance transcends national and sectoral boundaries, the Internet Governance Forum (IGF) was formed as a potentially effective and legitimate multi-stakeholder governance network that could better ensure cooperation. An evaluation of the impacts of the IGF model in national and regional contexts, particularly developing country contexts, has gained limited attention. This study is one of the first to explore the function of the IGF model in a regional context by investigating the impacts of the East African Community’s (EAC) East Africa Internet Governance Forum (EAIGF), the first regional IGF established globally.

Nonnecke, B., Aitamurto, T., Catterson, D., Crittenden, C., Garland, C., Huang, A., Krishnan, S., Nelimarkka, M., Newsom, G., Patel, J., Scott, J., & Goldberg, K. (2016). The California Report Card v1.0. In E. Gordon & P. Mihailidis (Eds.), Civic Media: Technology, Design, Practice (pp. 235 – 240). Cambridge, MA: MIT Press.



The California Report Card (CRC) v1.0 is an experimental platform developed by UC Berkeley in collaboration with the California Office of Lt. Gov. Gavin Newsom that streamlines public input by openly encouraging suggestions from a broad range of participants, and combining peer-to-peer review with statistical models to identify and highlight the most insightful ideas. The full case study is available at

Nonnecke, B.*, Krishnan, S.*, Patel, J., Zhou, M., Byaruhanga, L., Masinde, D., Meneses, M., Martin del Campo, A., Crittenden, C., & Goldberg, K. (October 2015). DevCAFE 1.0: A participatory platform for assessing development interventions in the field. Proceedings of the IEEE Global Humanitarian Technology Conference (GHTC). Seattle, WA.



The design and assessment of development initiatives is increasingly participatory, where decision makers consider feedback from affected populations. While digital data collection facilitates faster and more reliable analysis, existing data collection tools are not optimized for unstructured qualitative (textual) data and peer-to- peer participant collaboration. In this paper, we propose a system called the Development Collaborative Assessment and Feedback Engine version 1.0 (DevCAFE), a customizable participatory assessment platform that collects and integrates quantitative assessment, qualitative feedback and peer-to-peer collaborative filtering. DevCAFE incorporates a library of statistical analyses for researchers to quickly identify quantitative and qualitative trends while collecting field data. DevCAFE can run on any mobile device with a web-browser and can work with or without Internet connectivity. We present results from two pilot projects: (1) 137 participants evaluating family planning education trainings at three Nutrition Education Centers in rural Uganda, and (2) 4,518 participants evaluating policy priorities for elected leaders in the June 2015 Mexico mid-term elections. DevCAFE collected over 19,000 peer-to-peer ratings of 336 submitted ideas. Feedback gathered through DevCAFE enabled targeted reforms to the family planning efforts in Uganda and the need for increased government attention to public safety in Mexico. Case studies and interactive demos are available at:

Nonnecke, B., Krishnan, S., Akula, A., Cliff, A., Huang, A., Lin, A., Nehama, S., Byaruhanga, L., Masinde, D., Crittenden, C., & Goldberg, K. (May 2015). A mobile platform for participatory assessment of development programs and a case study in Uganda. Paper presented at the Humanitarian Technology Conference: Science, Systems and Global Impact. Boston, MA.



Development organizations need timely and reliable feedback on the efficacy of program interventions. Traditional assessment methodologies tend to be costly, time-consuming, and lack interaction: participants provide responses but seldom have access to the data or a role in its interpretation. We developed the Collaborative Assessment and Feedback Engine (CAFE) to engage communities from developing regions in collective assessment of local conditions, needs, and outcomes of development programs (see Figure 1).

The CAFE platform collects data in three phases: (1) Quantitative assessment, where participants submit responses on a scale from 0 to 10, “Strongly Disagree” to “Strongly Agree”, establishing baseline participant demographics and opinions. (2) Ideation, where participants provide a textual response to an open-ended question. (3) Collaborative evaluation, where participants evaluate the responses of others. CAFE provides researchers and participants with rapid, preliminary insights while in the field and can be used without Internet connectivity by taking advantage of ad-hoc networks (e.g., mobile devices networked to a laptop).

CAFE uses peer review to filter the most insightful ideas and Principal Component Analysis (PCA) to group participants based on their answers in the quantitative assessment phase. Insights from the statistical analysis are displayed to facilitate the discussion and incentivize participation. For example, median scores from the quantitative assessment and anonymous textual responses from the ideation phase are displayed immediately.

We present a case study on use of CAFE to assess family planning and reproductive health (FPRH) trainings at three health centers in Uganda in June 2014. Data were collected through CAFE from 137 women, including demographics, quantitative assessments of FPRH knowledge, and qualitative feedback on how to improve FPRH education strategies. Our data suggest that PCA can extract insights for further cross-sectional demographic analysis. For example, CAFE allowed us to identify that more tailored and locally relevant trainings are necessary to increase efficacy of FPRH outreach at each center. Our preliminary results suggest CAFE can collect reliable, direct and timely feedback on the efficacy of development programs.


Ochoa, M. & Nonnecke, B. (2019, October). Increasing human development in rural Mexico through policies for internet access. 2019 IEEE Global Humanitarian Tech Conference (GHTC). Seattle, WA.



Access and use of digital technologies is essential to ensure equal opportunities for education, employment, health, and political participation. Mexico is the 15th largest economy in the world, however there are important inequalities within the country that reduce its potential to reach development objectives. Only 20% of the rural population has internet access in comparison to 62% in urban areas. This digital divide threatens to further entrench social, economic, and political disparities. In this article, we analyze Mexico Conectado, a nationwide strategy to increase the digitalization of the country by creating internet access points in public spaces in rural and urban areas and supporting digital literacy training. We present a model to analyze the relationship and impact of Mexico Conectado on human development indicators in an index for the 32 States of Mexico. We found a positive relationship where rural internet access increases human development indicators measured in the United Nations Development Program’s Human Development Index (HDI). Our model shows that an increase of 1% in the number of households with internet access generates an increase of .02% in the HDI. Alternatively, if the number of households with internet access decreases by 1% the HDI would decrease by .07%. We conclude with policy recommendations for digital inclusion strategies to support human development.

Ochoa, M. & Nonnecke, B. (2019, September). Developing a Digital Inclusion Index at the State Level: The Case of Mexico. Paper presented at the 2019 Telecommunications Policy Research Conference (TPRC). Washington, DC.



Despite consensus over the importance of measuring digital inclusion, there is a lack of agreement among researchers on how it should be measured. This paper builds extensively off of prior research on measuring digital inclusion by analyzing different definitions for digital inclusion and comparing common digital inclusion indices. We analyze the variables included in five indices intended to measure digital inclusion: 1) Australian Digital Inclusion Index, 2) CISCO Country Digital Readiness, 3) ITU Digital Access Index, 4) The Economist Inclusive Internet Index, and 5) the World Bank Digital Adoption Index. Using a methodology called “qualitative meta synthesis,” we select variables that appear in at least three of the five original indices for inclusion in a more parsimonious index, the Digital Inclusion and Policy Index (DIP Index). We construct and analyze four versions of the DIP Index to validate fitness—two versions contain parsimonious combinations of variables from the five original indices and two versions add a variable to measure the presence of state-level digital inclusion policy. Data collected from the National Statistics and Geography Institute of Mexico for all 32 States of Mexico from 2018 are used to recreate four of the five original indices and our four versions of the DIP Index. These indices are correlated with each other and with an independent measure of economic and social competitiveness, the State Competitive Index developed by the Mexican Institute for Competitiveness. Theory supports a positive relationship between digital inclusion strategies and increasing state-level competitiveness in regard to economic and social factors. The results suggest that a parsimonious index of as few as four variables is capable of measuring digital inclusion and that an index of as few as five variables, including a variable to measure the presence of state level digital inclusion policy, is capable of measuring digital inclusion and impacts of state-level digital inclusion policy. The results create a parsimonious index for measuring digital inclusion, which can serve to alleviate costs for data gathering and elucidate the differing impacts of state-level policy strategies on digital inclusion.


Dinesen, B., Nonnecke, B., Lindeman, D., Toft, E., Kidholm, K., Jethwani, K., Young, H., Spindler, H., Oestergaard, C., Southard, J., Gutierrez, M., Anderson, N., Albert, N., Han, J., Nesbitt, T. (2016). Personalized telehealth in the future: A research agenda. Journal of Medical Internet Research, 8(3).



As telehealth plays an even greater role in global health care delivery, it will be increasingly important to develop a strong evidence base of successful, innovative telehealth solutions that can lead to scalable and sustainable telehealth programs. This paper has two aims: (1) to describe the challenges of promoting telehealth implementation to advance adoption and (2) to present a global research agenda for personalized telehealth within chronic disease management. Using evidence from the United States and the European Union, this paper provides a global overview of the current state of telehealth services and benefits, presents fundamental principles that must be addressed to advance the status quo, and provides a framework for current and future research initiatives within telehealth for personalized care, treatment, and prevention. A broad, multinational research agenda can provide a uniform framework for identifying and rapidly replicating best practices, while concurrently fostering global collaboration in the development and rigorous testing of new and emerging telehealth technologies. In this paper, the members of the Transatlantic Telehealth Research Network offer a 12-point research agenda for future telehealth applications within chronic disease management.


Leiken, Emma. (2023). TikTok and a Broader Policy Agenda. Tech Policy Press.



Written by 2023 Tech Policy Fellow Emma Leiken. The article discusses the rise of TikTok as a global social media platform and its implications for policy and regulation. She argues that TikTok’s success highlights the importance of creating a comprehensive policy framework that addresses issues such as data privacy, content moderation, and competition in the digital marketplace. She examines the challenges faced by governments and policymakers in regulating social media platforms like TikTok, and suggests a range of policy responses.

Farid, Hany and Nonnecke, Brandie. (2023). The Case for Regulating Platform Design. Wired.



The article discusses Section 230, a law that shields internet platforms from legal responsibility for user-generated content. The authors argue that the law should be reformed to incentivize big tech companies to prioritize user safety in their design choices. The authors propose a new regulatory approach that would hold platforms accountable through a focus on design choices and product liability instead. The approach would encourage companies to invest in safer design and moderation practices and could reduce the spread of harmful content online. The article also discusses the potential challenges and criticisms of this approach, including concerns around free speech and the difficulty of regulating complex algorithms.