Responsible Machine Learning
Advancing the responsible use of algorithmic decision-making systems
#ResponsibleML is Twitter’s guide for positive change.
The #ResponsibleML program at Twitter guides our machine learning practices and shapes how we apply algorithms in responsible and ethical ways. We’re action-oriented and apply learnings from our research to create engineering infrastructure and tools to continually improve the experience on Twitter.
Our #ResponsibleML program will:
Communicate openly to further ResponsibleML.
• Aim to increase public understanding around RML as an emerging field and its nuanced issues
Research. Learn. Apply. Share.
• Share research, results, how we apply them, and what we learn as a result
• Develop and share best practices and tools
• Share how we attempt to solve hard problems
• Examine, shine light on, and work to reduce AI bias—and then communicate about it in the open
• Influence and shape more ethical product, design, and engineering choices
Empower & Educate.
• Thoughtful and clear communication on how machine learning shapes the experience on Twitter
• Identify, improve, and explain how models make decisions
• Empower customers with tools, control, and education to deepen their understanding
Why
Twitter serves the public conversation with billions of people talking each day. We seek to improve social media’s impact on the world and make Twitter an example of algorithmic ethics done right. We’re creating a space for free expression and building a culture of trust and respect as we apply the meaningful work we are doing. We’re committed to promoting online #health, earning #trust, and doing so in a #straightforward and transparent way.
The difference
We’re pioneering engineering and research in the field of ML within Twitter and the industry. We are an engineering team and a research team, which is what makes us unique. Our work is actionable, and we apply learnings from our research to improve the equity and fairness of our product and inform the way we design and build our service.
Our approach
Transparency. When we share our findings openly, it benefits the #ResponsibleML community as a whole. In doing so, this helps the evolving field of machine learning develop ethical, responsible, inclusive, and outward-looking practices. Our goal is to be a leader in the field of responsible machine learning, guide the development of our products, and benefit our customers with this approach to transparency and accountability.
Machine Learning, Ethics and Transparency (META)
The Machine Learning, Ethics, Transparency and Accountability (META) team is a part of Twitter’s machine learning discipline. We aim to create easy-to-use tools and utilize well-established investigation methodologies and frameworks to understand and mitigate the bias in our products powered by machine learning.
The focus is on ensuring that Twitter’s engineers, data scientists, and researchers address the societal impact of the ML software they build with a thoughtful and insightful approach to the potential ways our end customers may be impacted.
Featured blog posts
Join us
We’re searching for smart and curious machine-learning researchers, data scientists, product managers, software engineers and more who want to be part of a world‑class team working on the most exciting technology applied to Twitter data.