How do you report the social impact of your algorithms to clients?

1 Define social impact

Before you can report the social impact of your algorithms, you need to define what it means and how you measure it. Social impact is the positive or negative change that your algorithms create or contribute to in the lives of individuals, communities, or the environment. It can be quantified by indicators such as fairness, accuracy, transparency, accountability, privacy, or sustainability. You should align your definition and measurement of social impact with your client's goals, values, and expectations, as well as with relevant standards and frameworks.

2 Identify stakeholders

Another key step for reporting the social impact of your algorithms is to identify who are the stakeholders that are affected by or interested in your work. Stakeholders can include your client, their customers, their employees, their competitors, their regulators, or the general public. You should map out the stakeholder groups, their needs, their concerns, and their influence on your algorithm development and deployment. You should also engage with them to solicit their feedback, input, or participation in your social impact assessment.

4 Highlight benefits

One of the main objectives of reporting the social impact of your algorithms is to highlight the benefits that they bring to your client and their stakeholders. Benefits are the positive outcomes or value that your algorithms deliver or enable, such as improving quality, reducing costs, increasing productivity, or enhancing customer satisfaction. You should showcase the benefits of your algorithms by using concrete examples, testimonials, case studies, or metrics. You should also relate the benefits to the social impact indicators that you defined and measured.

5 Address risks

Another important objective of reporting the social impact of your algorithms is to address the risks that they pose or face. Risks are the potential negative consequences or threats that your algorithms may cause or encounter, such as errors, biases, discrimination, fraud, or misuse. You should acknowledge the risks of your algorithms by using scenarios, simulations, audits, or tests. You should also demonstrate how you mitigate or manage the risks by using methods such as validation, verification, monitoring, or remediation.

6 Provide recommendations

The final step for reporting the social impact of your algorithms is to provide recommendations for your client and their stakeholders. Recommendations are the actions or suggestions that you propose to improve, maintain, or optimize the social impact of your algorithms. They can include technical, operational, or strategic aspects, such as updating, refining, or scaling your algorithms, or aligning them with best practices, policies, or regulations. You should justify your recommendations by using evidence, logic, or benchmarks, and indicate the expected outcomes, costs, or benefits of implementing them.

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?