1 Understand the purpose and context of your image analysis

Before you collect, process, or share any images, you should have a clear and legitimate purpose and context for your image analysis. For example, are you using it for research, education, entertainment, security, or marketing? How will your image analysis benefit or harm the individuals or groups in the images? What are the potential risks and benefits of your image analysis for society? You should also consider the cultural, social, and legal norms and expectations of the people and places in the images, and respect their dignity and autonomy.

2 Obtain informed and explicit consent from the data subjects

One of the most important principles of ethical image analysis is to obtain informed and explicit consent from the data subjects, that is, the people who are depicted or identifiable in the images. This means that you should inform them about the purpose, scope, methods, and outcomes of your image analysis, and obtain their voluntary and unambiguous agreement to participate. You should also provide them with the option to withdraw their consent at any time, and to access, correct, or delete their data. You should avoid using images that are obtained without consent, or that are obtained under coercion, deception, or manipulation.

3 Protect the privacy and security of the data

Another key principle of ethical image analysis is to protect the privacy and security of the data, that is, the images and the information derived from them. This means that you should apply appropriate technical and organizational measures to prevent unauthorized access, use, disclosure, or modification of the data. For example, you should encrypt, anonymize, or pseudonymize the data, limit the access and retention of the data, and monitor and audit the data processing activities. You should also comply with the relevant data protection laws and regulations, such as the General Data Protection Regulation (GDPR) in the European Union.

4 Minimize the bias and error in the image analysis

Image analysis is not infallible, and it can produce biased or erroneous results, depending on the quality, quantity, and diversity of the data and the algorithms. These results can have negative consequences for the data subjects and the society, such as discrimination, exclusion, or injustice. Therefore, you should minimize the bias and error in the image analysis, by ensuring that your data and algorithms are representative, accurate, reliable, and transparent. You should also test and validate your image analysis results, and disclose any limitations or uncertainties.

5 Respect the intellectual property and copyright of the images

Image analysis often involves using images that are created, owned, or licensed by other parties, such as photographers, artists, or publishers. You should respect the intellectual property and copyright of the images, and obtain the necessary permissions or licenses to use them for your image analysis. You should also acknowledge the sources and authors of the images, and avoid infringing or misappropriating their rights.

6 Communicate and report your image analysis responsibly

Finally, you should communicate and report your image analysis responsibly, that is, in a way that is honest, accurate, clear, and respectful. You should explain the purpose, methods, results, and implications of your image analysis, and avoid exaggerating, misleading, or manipulating the audience. You should also respect the privacy and dignity of the data subjects, and avoid disclosing or publishing any personal or sensitive data without their consent. You should also seek feedback and input from the relevant stakeholders, such as the data subjects, the image owners, or the image analysis users.

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?