RESEARCH

ML APPLICATIONS

Introducing Open Loop, a global program bridging tech and policy innovation

January 11, 2021

The rapid advance of emerging technologies such AI can make it difficult to fully understand and anticipate how they might eventually impact communities around the world. To help develop forward-looking policies around the development and use of new technology, Facebook is launching Open Loop, a global strategic initiative and consortium to connect policymakers and technology companies to develop evidence-based policy recommendations.

Open Loop builds on Facebook’s ongoing work to help develop policy recommendations around the use of technology. The initiative will prototype and test new and different approaches to laws and regulations before they are enacted, so stakeholders can better understand how well they’ll work in the real world.

Something went wrong
We're having trouble with playing this video.

How Open Loop works

As the name suggests, Open Loop’s goal is to create a robust collaborative feedback loop (an “open loop”) of practical learnings between the people who make policy and those who have to implement it. As part of Open Loop, policy makers will be able to work with a vibrant community of tech companies, including Facebook, to build sound and operational governance frameworks, and discuss regulatory best practices.

Participating companies will be able to leverage training, tutorials, toolkits, mentorship, and technical assistance provided by the consortium, while sharing practical insights and working directly with policy makers to inform new regulatory solutions. Open Loop will take an experimental, interactive approach similar to how technology is built: alpha phases to research and test different regulatory pathways; and beta phases to iterate and refine these frameworks before sharing them more broadly.

Open Loop will follow four main steps when rolling out its programs. First, the consortium will create policy prototypes on specific topics related to emerging technologies like AI. Second, Open Loop will test and evaluate these prototypes under real world conditions by collecting information from participants as they apply these frameworks to their specific products and services. By asking questions about how clear, applicable, and effective these policy prototypes are, Open Loop collaborators will learn about their effects, strengths, and limitations. Third, Open Loop will apply the lessons learned so participants can iterate on and improve these normative frameworks. And fourth, Open Loop will deliver evidence-based policy recommendations to policy makers based on the findings of the program and the feedback collected.

Policy prototyping underway in Asia and Europe

As part of the Open Loop initiative, Facebook and its partners have begun to roll out a series of policy prototyping programs. In Asia, we are collaborating with Singapore InfoComm Media Development Authority (IMDA) and Personal Data Protection Commission (PDPC), and with 12 AI companies from the Asia-Pacific region, to test specific concepts, processes, and guidance on AI explainability and transparency. These are based on Singapore’s Model AI Governance Framework and its companion guide – the Implementation and Self-Assessment Guide for Organizations.

In Europe we have partnered with 10 European AI businesses to co-create and test an AI Risk Assessment framework, akin to GDPR’s Data Protection Impact Assessments (DPIAs), on different AI applications. The results and recommendations of this program are being published today in the first Open Loop report, “AI Impact Assessment: A Policy Prototyping Experiment.”

As today’s report outlines, the results of Open Loop’s first policy prototyping experiment were very promising. Based on feedback from the companies we collaborated with, our prototype version of a law requiring AI risk assessments, combined with a playbook for how to implement it, was clearly valuable to the participants as a tool for identifying and mitigating risks from their AI applications that they may not have addressed otherwise.

The experiences of our partners highlighted how this sort of risk assessment approach can inform a more flexible, practicable, and innovative method to assessing and managing AI risks compared to more prescriptive policy approaches.

We invite policy makers to join us in supporting or participating in experimental policy programs like these, and to test novel regulatory approaches to complex technology policy issues before codifying them in law. Ultimately, we hope Open Loop will help shape and inform public policies around emerging technologies and help the tech industry build products and services that benefit communities around the world.

What startups are saying about Open Loop

As AI regulation advances, it is critical that innovative companies of all sizes, especially startups, have their voices heard in these discussions. In doing so, we can help ensure that regulation is clearer, more inclusive, achieves its goals, and fosters greater innovation and competition.

The participation of several Europe-based AI startups has been instrumental in launching Open Loop. Here’s what some of them are saying about the value of this program:

  • EVO Pricing: “The policy prototyping launched by Facebook has fostered a healthy debate at Evo Pricing about the future AI governance questions that span across 3-5 years beyond our typical 12 months planning horizon. Governance is a key topic for us at Evo Pricing, as every day we collect and process data on over 1.3 billion people and over 900 million products globally. Facebook has pushed our thinking into how to prototype a policy document that will make our autonomous supply chain future-proof, while also nudging the public agenda and discourse further.”

  • RiAtlas: “For a startup like RiAtlas, the participation in Facebook’s ADIA Policy Prototyping Program was an excellent opportunity to understand and mitigate in advance the AI-tools risks in digital health. At RiAtlas, transparency and accountability requirements raise the chance to innovate responsibly by applying AI to digital health. Today, our stakeholders will be fully aware of the confidence and explanation of the AI-tool conclusions provided by RiAtlas.”

  • Irida Labs: “Our participation in the Policy Prototyping Program gave us the opportunity to participate at the AI Governance dialogue in Europe and support the European AI Strategy and the Coordinated Plan. In Irida Labs’ daily practice the role of AI and the responsible use of data was always of paramount importance. Thus, our participation provided us the chance to reflect and reassess all internal AI /ML development processes in terms of safety and liability, which are the core of Irida Labs’ vision intelligence solutions. We would like to thank Facebook for their invitation, the program facilitation, and their commitment towards a Responsible AI.”

  • Allegro AI: “Allegro AI is proud to have been a participant in the AI Policy Prototyping program initiated by Facebook. As a leading ML-Ops tool provider we support thousands of customers who deal with building AI models on a daily basis. Making sure these models do not include inherent biases is crucial to the success of their products and services. We believe this initiative is a great step to set best practices in the field.”

  • NAIX: “A high standard of risk assessment and governance is of utmost importance for NAIX as we handle sensitive personal identifiable information. The Facebook Policy Prototyping Program has given NAIX the opportunity to converse with policy makers and other leading industry partners, and thereby enabled us to refine and adapt our AI application governance in many aspects. This helped us ensure that we continuously maintain our high standards and we are grateful to have been given the opportunity to participate in building a first-of-its-kind AI policy prototype with Facebook and its engaging network.”

  • Rogervoice: “The Facebook [Policy Prototyping Program] program has helped us to shape our data privacy issues especially as regards AI development. This is especially critical as we handle the largest global telecommunications service for the deaf and hard-of-hearing. We take policy development very seriously and hope that our user's interest is best represented.”

Written By

Norberto Andrade

Global Policy Lead for Digital and AI Ethics