CompanyStack Overflow AdvertisingStack Overflow for Teams
Contact legal Submit Data Request

Consolidated Responsible AI policy

This policy sets out the approach to responsible AI deployment.

Last updated

1. Objectives

This policy sets out the approach to responsible AI deployment across the Prosus Group (“Prosus”) and Stack Overflow (“Stack,” “we,” “us”).

The objectives of this policy are:

  • To ensure that ethical considerations are an integral part of AI development across Stack Overflow, as this is essential to preserve quality and longevity in AI products and their performance, as well as to maintain public trust.
  • To ensure data science and AI add social value for consumers, partners, and Stack, and at the same time apply globally-benchmarked responsible AI standards, whether set forth in applicable laws, ethical best practices, or voluntary frameworks implemented in relevant jurisdictions.

2. Accountability

Stack Overflow is responsible for implementing this policy and should designate appropriate stakeholders with responsibility for AI governance within the relevant business. Stack is responsible for meeting AI objectives, providing guidance, and should designate appropriate stakeholders with responsibility for AI governance.

3. Responsible AI guiding principles

Because we recognize that achieving state of the art ethical and responsible AI is a process that never concludes, we strive to develop strong capabilities in this area. The guiding principles we use to inform this ongoing work are:

  • Force for good. We are committed to ensuring that our AI applications contribute to positive changes in people's everyday lives and are in line with our vision as a group to invest in and build companies that empower people and enrich communities across the world. In this way, we aim for our AI endeavors to benefit our customers, our company and our stakeholders directly, (whilst creating benefits for society at large).
  • Technical excellence and robustness. We strive to ensure that the scientific and technical standards informing research, design and application of our AI products and services are sound, robust and on par with the global best practices.
  • Fairness and inclusivity. We strive to avoid creating unfair biases and reinforcing or exacerbating existing biases. We ensure that fairness and inclusivity considerations are appropriately incorporated in the development and deployment of AI systems.
  • Accountability & transparency. The designers of systems have an ongoing duty to appropriately document how they implement these principles, and to monitor and control for performance over time. Furthermore, AI systems and their outputs should be context-sensitive, transparent and explainable to different stakeholders in appropriate circumstances. Where possible, we disclose relevant information regarding the use of AI in non-technical terms to foster trust from individuals and businesses alike.
  • Privacy & security. Data used to train and deploy AI systems must be proportionate, handled in conformity with applicable data protection laws. We make sure that the AI we develop, use and deploy are designed in a way that minimizes known security threats, through appropriate technical and organizational measures.

4. Responsible AI governance operating framework

Our responsible AI operating framework aims to ensure that social and ethical dimensions of AI are integrally included within the product or feature development process, organized around four key pillars that directly apply Stack Overflow:

Govern

Stack will specify and announce the governance principles applicable to their own development and deployment of fair and responsible AI in a manner that is reasonably accessible by their users, clients, partners and the public. This might include, for example, publishing a policy compatible with the standards contained in this Prosus Group Policy on Responsible AI on their own websites. Such disclosure should always be adapted to reflect the priorities and focus of the individual companies and should be developed in line with appropriate transparency benchmarks, including this policy.

Should Stack operate in jurisdictions where specific AI laws apply, including but not limited to the European Union Artificial Intelligence Act (EU AI Act), they must ensure that the requirements of such laws are met.

Stack has evaluated its products and services under the EU AI Act and remains dedicated to maintaining compliance, such as including required disclosures about AI use.

Design

Stack will appropriately design for privacy, security, transparency, bias controls, and robustness as an integral part of development and deployment of AI systems and models. This pillar includes pursuing efforts aimed at introducing more explainable and robust models.

Monitor

Stack will implement appropriate processes aimed at auditing for accountability, bias and risks implicated by specific AI models. This can be enabled through adopting tools for bias check as part of model-development practices. Proactive cooperation to monitor across disciplines also helps to preempt AI-related risks, analyze the likelihood of harm and to appropriately mitigate for such risks.

Train

Stack will aim to increase AI literacy, by creating and implementing appropriate AI training programs. In this context, all Stack employees are expected to become well-acquainted with company AI tools and to actively use them as aids in the execution of their work, to drive greater scale, efficiency, and quality of outputs.

5. AI and investing

As investing in innovative business models and AI-first companies is part of our investment strategy, we also consider the principles contained in this Policy in the context of our acquisitions, investments, and partnerships.

6. Group support and monitoring

The Prosus Group’s AI Team, in collaboration with its AI & Ethics Working Group, helps to support businesses with their responsible AI implementations, legal compliance efforts, regulatory engagements, AI risk assessments and any AI implications of mergers and acquisitions activity.

The Prosus Group AI team provides highly specialized training on several AI themes, including responsible AI. It organizes technical and scientific workshops for the community, connects data scientists working on similar initiatives, shares practices, tools and lessons learned across businesses.

To support its board members, executives and management teams, Prosus ensures that responsible and ethical AI is a component of the training offered in our AI4Impact leadership training series and future trainings deployed through the Prosus Academy.

The Risk Committee of the Prosus Board of Directors reviews this policy and its implementation on an annual basis, as part of its oversight and governance responsibilities.

End-users, consumers and other stakeholders that want to address queries or report concerns related to the implementation of this Policy can reach out to Prosus and/or group companies through the communication channels they make available.

For more information about this policy, please contact privacy@prosus.com.