Green Consumption Assistant
- Berlin, Germany
- January 2021
- Ecosia
The Green Consumption Assistant (GCA) supports sustainable consumption on Ecosia. The GCA recommends green products and provides information about more sustainable alternatives, for example in the form of hints for repair, lending, or sharing options. The basis for the GCA's recommendations is a product database (GreenDB) with ecological and social sustainability information developed with the help of machine learning methods.
The GCA is a joint project between the TU Berlin, the Berliner Hochschule für Technik, and the green search engine Ecosia and is funded by the German Federal Ministry for the Environment as a lighthouse project for artificial intelligence.
- Open source
- Public
Project stage (in ):
Implemented/Running Project
Implemented by:
Inhouse
Industrial Sectors:
Information and communication, Professional, scientific and technical activities
Usage of AI:
Natural Language Processing, Data Management and Analysis, Human-Computer Interaction, Information Retrieval
Generation of AI:
Traditional Machine Learning (Linear Regression, CART, SVM, etc.), Deep Learning (CNN, Transformers, etc.)
Model training:
Supervised Learning
Motivation and values
How did the idea of your project come about?
Internal in our organization
What is the goal of your project in relation to the public interest?
The production, shipping, usage, and disposal of consumer goods have a substantial impact on greenhouse gas emissions and the depletion of resources.
Modern retail platforms rely heavily on Machine Learning (ML) for their search and recommender systems. Thus, ML can help to foster sustainable consumption patterns by accounting for sustainability aspects in product search or recommendations. The lack of large high quality publicly available product data with trustworthy sustainability information impedes the development of ML technology that can help to reach our sustainability goals. The GreenDB collects on a weekly basis sustainability information and credibility ratings of products in European online shops.
Importantly we prioritize which products anonymized users care about most, based on search logs of millions of users.
Product categories are prioritized using search logs of millions of users, and sustainability information are enriched with credibility information by human experts.
Hence our data set enables researchers to train ML models that can automatically differentiate trustworthy sustainability information from greenwashing for product categories users care most about.
We present initial results demonstrating that ML models trained with our data can reliably predict the sustainability label of products. These contributions can help to complement existing e-commerce experiences and ultimately encourage users to more sustainable consumption patterns.
Design & Safeguards
Which stakeholders were involved in the process of development and implementation of the project?
Developer, External domain expert, Public administration, Academic researcher, Industry partners, Product designers, users
How did you engage relevant stakeholders?
Workshops, Survey, Interviews, Co-creation process
Did you apply specific methods of participatory design, and if yes, which ones?
We involve a wide range of stakeholders through formats ranging from interviews, workshops, and survey, to live co-creation sessions.
Have the project results been validated by third parties?
Yes
Have the design and the results of the project been made transparent to the public?
Yes
In what way have the design and the results of the project been made transparent to the public?
By academic researchers
Involving the people who will be affected by the project is a necessary part of the project design
5
Which direct channels for feedback exist?
Users are encouraged to give feedback directly on the product via feedback links and can use various other platforms to give feedback.
How did you verify the accuracy and robustness of your system?
Feedback including both qualitative and quantitative surveys and interviews and anonymized tracking data
Which transparency measures do you use to document your use of data and possible biases or limitations of the data sets?
Data sheets
What technical and organizational safeguards have you implemented to protect personal data and mitigate possible harms? Choose all that apply
Data minimization (incl. not gathering personal data), Data aggregation and summarization, De-identification (incl. obfuscation), Maintaining an up to date privacy policy
Do you take measures in regards to ecological sustainability of your project? Please elaborate.
We leverage state of the art model selection techniques to ensure not only optimal performance of our models in terms of held-out-data-set metrics but also to minimize energy consumption.
Is your project directly or indirectly contributing to solutions for the climate crisis or other UN sustainability goals?
Indirectly
Please elaborate
By fostering sustainable consumption, the project directly contributes to SDG 12 (Responsible Consumption and Production).
Does your project rely partly or fully on the use of open data and/or publish results in an open data set? Please elaborate.
All our research artefacts, code and data, is fully open sourced and available for research.