Sama
A social-impact technology company providing data annotation, validation, and training-data services for artificial intelligence.
Last updated August 31, 2026
AI crawler reads
No AI crawler has read this profile since Oct 28, 2026.
Overview
Sama, originally known as Samasource, is a San Francisco-based technology and impact-sourcing company that supplies annotated and validated data for artificial-intelligence and machine-learning systems. Its work covers image, video, sensor, and other forms of data used to train or evaluate computer-vision, natural-language, autonomous-navigation, augmented-reality, biotechnology, agriculture, manufacturing, and e-commerce applications. The company has operated internationally, with a particularly significant presence in Africa. Leila Janah founded the organization in 2008. Its original model combined technology outsourcing with a poverty-alleviation mission: complex digital projects were divided into smaller tasks that could be completed by workers in developing countries after relatively short periods of training. Janah used the term “microwork” to describe this approach, presenting online employment as a means of connecting people with limited local opportunities to global technology work. Sama’s commercial service model is built around a secured cloud-based annotation platform. A typical project can involve uploading source data, assigning annotation tasks, sampling and validating results, applying quality assurance procedures, delivering completed datasets, and coordinating collaboration between clients and distributed production teams. The resulting datasets may be used to train machine-learning models, measure model performance, or improve the ability of automated systems to recognize objects, actions, language, or potentially harmful content. The organization began as a nonprofit and later developed a hybrid structure. In 2019, it moved toward a for-profit business model while the earlier nonprofit entity became a shareholder. This change reflected the company’s effort to combine commercial demand for high-volume training data with its original social-enterprise and impact-sourcing objectives. Sama has attracted both commercial attention and criticism. Its services have been associated with content-moderation and artificial-intelligence safety work in which workers reviewed disturbing material, including violence, hate, sexual abuse, and other forms of harmful content. Reporting and subsequent legal proceedings raised questions about recruitment disclosures, pay, psychological support, labor rights, privacy, and working conditions, particularly among workers in Kenya. The company has denied wrongdoing and has stated that it maintains transparent hiring practices and prioritizes worker health and well-being. The controversies have placed Sama within broader debates about the hidden human labor behind AI systems and the tension between socially oriented outsourcing and the conditions under which that work is performed.
History
Sama was founded in 2008 by entrepreneur Leila Janah under the name Samasource. Janah developed the organization around the idea that internet-enabled “microwork” could connect people in poorer regions with digital employment generated by companies in wealthier markets. Rather than treating outsourcing solely as a cost-reduction mechanism, the original model sought to direct work toward women and young people in developing countries and to provide training that would allow them to complete basic digital tasks. The company’s early identity was that of a nonprofit or socially oriented enterprise. Its operational approach involved dividing large and technically complex data projects into small, standardized assignments. Workers could complete these assignments after training in basic English, task procedures, and quality requirements. This structure made it possible to process large volumes of material while distributing work across international production centers. As demand for artificial intelligence expanded, Sama increasingly focused on the data infrastructure required to build and evaluate machine-learning systems. The company developed services for annotating images, video, and sensor data, as well as validating labels and maintaining quality-control processes. These services became relevant to sectors such as automotive technology, navigation, augmented and virtual reality, biotechnology, agriculture, manufacturing, and online commerce. A secured cloud platform supported the project lifecycle, including data ingestion, task allocation, annotation, sampling, review, quality assurance, delivery, and client collaboration. The company’s social mission and commercial activities evolved together. In 2019, Sama adopted a hybrid business structure, becoming a for-profit company while the former nonprofit organization became a shareholder. The restructuring enabled the business to pursue commercial contracts while retaining a formal connection to its impact-oriented origins. The company continued to describe its work as a way of creating opportunities in communities with limited access to conventional employment. Sama later became associated with content moderation and AI-safety projects. In one widely reported arrangement, Kenyan workers reviewed harmful or toxic material used to develop systems intended to identify abusive, violent, hateful, or otherwise unsafe content. These assignments generated criticism because workers reported exposure to disturbing material and alleged that recruitment, pay, health support, and working conditions were inadequate. The controversy highlighted a structural issue in the AI industry: systems presented as automated often depend on large amounts of human judgment performed by contractors and subcontractors. Labor disputes intensified in Kenya. Former employees and their representatives raised concerns involving labor and privacy law, healthcare and insurance, compensation, unionization, alleged retaliation, and the psychological consequences of moderation work. Former content moderators subsequently brought claims against Sama and, in some cases, related client companies. Sama denied wrongdoing and maintained that its hiring process was transparent and that employee health and welfare were priorities. The disputes, together with external assessments of working conditions, contributed to sustained scrutiny of the company’s impact-sourcing model. In 2023, Sama workers participated in the creation of the African Content Moderators Union, joining a wider effort by moderation workers to obtain collective representation. Sama’s history therefore spans two connected but sometimes conflicting narratives: the use of digital work to widen access to employment, and the risk that globally distributed AI labor can reproduce low pay, limited bargaining power, privacy concerns, and psychological harm. The company remains known primarily as a provider of training data and data-annotation services, with operations extending beyond the United States.
- 2023
Participation in African Content Moderators Union
[1]Sama employees took part in establishing a union representing content moderators in Africa.
- 2022
Public scrutiny of content-moderation work
[1]Investigations and legal claims brought attention to the conditions experienced by Kenyan workers reviewing disturbing material for AI and online-safety projects.
- 2019
Adoption of a hybrid business model
[1]The organization shifted toward a for-profit structure, while the previous nonprofit entity became a shareholder.
- 2008
Samasource founded
[1]Leila Janah founded Samasource as a socially oriented organization focused on connecting people in developing countries with digitally delivered work.
Products and positioning
Social enterprise and impact-sourcing provider for AI data annotation and machine-learning services
AI training data
Image annotation
Annotation of images for computer-vision systems, including labeling objects, scenes, attributes, boundaries, and other visual features. The resulting datasets can support model training, validation, and performance measurement across industries such as automotive, navigation, retail, agriculture, and manufacturing.
AI training data
Video annotation
Review and labeling of video sequences for machine-learning applications. Video work can capture objects, actions, movement, events, and changes over time, providing structured examples for systems used in computer vision, safety, navigation, and augmented or virtual reality.
AI training data
Sensor data annotation
Processing and validation of sensor-derived data for systems that interpret physical environments. Such work is relevant to navigation, automotive technology, robotics, and other applications that depend on labeled or quality-checked machine-learning inputs.
Data services
Data validation and quality assurance
Quality-control services covering sampling, verification, review, and delivery of annotated datasets. These processes are intended to improve label consistency and provide clients with data suitable for training, testing, and evaluating AI models.
Controversies
2023
Labor and privacy litigation
Former workers and their representatives pursued claims concerning labor rights, privacy, compensation, health protections, alleged union-busting, and psychological harm connected with moderation and AI-data work.
2022
Content-moderation worker conditions
Reporting and legal claims alleged that Kenyan moderators were exposed to traumatic material under conditions involving low pay, insufficient psychological support, recruitment concerns, and weak labor protections. Sama denied wrongdoing and said its hiring practices were transparent and its workplace culture prioritized employee well-being.
Brand decisions
2019
Transition to a hybrid commercial structureStrategyBackground
Samasource had begun as a nonprofit-oriented organization but was operating in a growing commercial market for AI training data and annotation.
Action taken
The organization adopted a for-profit business model while the earlier nonprofit entity became a shareholder.
Aftermath
The change allowed Sama to pursue commercial data-services work while maintaining a formal relationship with its social-impact origins.
Recent events
- 2023
African Content Moderators Union formed with participation from Sama workers[1]
Other
Sama employees participated in the formation of the African Content Moderators Union alongside workers from other outsourcing companies, reflecting broader organizing efforts among workers performing online safety and moderation tasks.
- 2023
African Content Moderators Union formed[1]
Other · Leadership change
Sama employees joined content moderators from other African outsourcing companies in forming the African Content Moderators Union.
Sources1 sources
Cite this profile: Cite the canonical profile. /brand-wiki/samasource · Editorial policy · How profiles are compiled