Omdena Code of Ethics: AI is for Everyone

Table of Contents

• Introduction
• AI Built on the Three C’s is Essential
• Collaborator Honor Code
• Data, IP, and Confidentiality
• Transparency and Explainability
• Ethical AI Decision-Making Framework
• Bias Mitigation Strategies
• Community Engagement
• Responsibility of Policymakers
• Continuous Learning and Ethical Training
• Ethical AI Auditing and Monitoring
• Reporting Mechanism

1. Introduction

At Omdena, we’re all about shaping the future of AI in an ethical, inclusive, and democratic way. Our Code of Ethics lays out the principles that guide AI developers, organizations, policymakers, and everyone who’s passionate about responsible AI solutions.

2. AI Built on the Three C’s is Essential 

Omdena believes that the best bet we have to tackle ‘bad’ AI is we the people. Collaboration among varied talents enables us to bridge gaps in understanding between different mindsets, share knowledge, and unite people and values. It, therefore, helps to create compassion and harnesses crowd wisdom, diversity, and inclusion to serve the long-term interests of those communities. 

The other key element is consciousness. As so much division exists in this world, we need to understand that deep down we all are one. Thus our consciousness is collective. Through forming a sense of community, we collaborate together with compassion and consciousness. Thus, AI built by the three C’s (Collaboration, Compassion, and Consciousness) will help us to remove endemic sociological and historical bias and other inequalities that exist in society. 

Omdena’s framework of bottom-up collaboration helps to achieve AI models built with the three C’s.

3. Collaborator Honor Code

Let’s break down Omdena’s Honor Code into something simpler:

• Originality: We play by the rules and avoid copying others’ work. Plagiarism doesn’t sit well with us, especially in the world of AI.
• Plagiarism Reporting: If you see someone playing the copycat, let us know. We’re all in this together to keep things ethical.
• Respectful Behaviour: Politeness is the name of the game here. We don’t do harassment or personal attacks. It’s all about constructive conversations.
• Privacy and Diversity: We respect your privacy and diversity. No sneaky data collection without permission. We love open-minded discussions.
• Appropriate Content: Keep it respectful, non-invasive, and copyright-friendly. Spamming is a no-go.
• Direct Communication: Use direct messages wisely. Don’t overdo it unless it’s for collaboration. And talking to partner organizations? Check with us first.

4. Data, IP, and Confidentiality

We’ve got your data and intellectual property covered:

• Data Protection: No unauthorized data sharing here. We follow the rules and ethics.
• Compliance: We respect data agreements and licensing conditions. Open-source data sets are our friends.
• Open-Source: Our open-source code lives in our GitHub repository, and we share it with the world under the GPL 3 license.
• IP and Code Ownership: For our proprietary projects, we maintain a range of ownership arrangements, ranging from joint intellectual property ownership to complete ownership by our associates and clients.
• Data and Code Security: We ensure data and code security by working on client and cloud infrastructures, enforcing NDAs, applying data anonymization, and using advanced security tools and protocols. 

5. Transparency and Explainability

We’re all about transparency. Not only do we document our models and decision-making processes, but we also employ model explainability techniques such as LIME, SHAP, and feature importance visualization to demystify our algorithms.  

The data we utilize originates from their expertise, and our models are designed to provide understandable outputs, serving as valuable assets to augment their work rather than replace it. 

6. Ethical AI Decision-Making Framework

Ethics matter. We’ve got a framework to help us make ethical AI decisions during projects. Our ethics approach is built into the design phase of each closed-source project. 

Our ethical framework is based on the following principles:

• Transparency and explainability: We make our AI systems, as much as possible, transparent and explainable so that everyone can understand how they work. This helps to build trust and rapport with stakeholders.
• Fairness: We consider the potential impact of our AI systems on different groups of people and take steps to mitigate bias. We want our AI systems to be fair and equitable for everyone.
• Privacy: We respect the privacy of individuals whose data is used in our AI projects. We obtain informed consent and use data in a way that is respectful and responsible.
• Data quality: We ensure that the data used to train and test our AI models is accurate and representative of the real world. This helps to avoid bias and ensure that our AI systems are making decisions that are in line with reality.
• Accountability: We define clear roles and responsibilities for all stakeholders involved in our AI projects. This helps to ensure that everyone is held accountable for their actions and that our AI systems are used in a responsible way.
• Beneficence: We design our AI systems to benefit humanity and do not cause harm. We consider the potential negative impacts of our AI systems and take steps to mitigate them.
• Public engagement: We involve the public and relevant stakeholders in the decision-making process for our AI projects. We want to ensure that our AI systems are developed in a way that is responsive to the needs of the people who will be using them.
• Continuous monitoring and evaluation: If asked for, we can support a client to regularly monitor and evaluate the performance of AI systems to ensure that they are being used in an ethical way. This allows us to identify any potential problems early on and take steps to address them.
• Compliance with regulations: We try to ensure that our AI projects comply with all applicable laws and regulations. This helps to avoid legal problems and ensure that our AI systems are being used in a responsible way.
• Education and training: We provide education and training to team members and stakeholders on ethical considerations in AI and data science. We want to ensure that everyone involved in our AI projects understands the ethical issues involved and is committed to using AI in a responsible way.
• Sustainability: We assess the environmental impact of our AI projects and strive to minimize it. We want to ensure that AI is used in a way that is sustainable for the planet.
• Community and collaboration: We encourage collaboration with other organizations and researchers to advance ethical practices in AI and data science. We want to share knowledge and best practices and build a more ethical AI community.

7. Bias Mitigation Strategies

We’ve got strategies to keep our AI algorithms fair and unbiased. It is important to note that there is no single silver bullet for mitigating bias in AI systems. A combination of the strategies is often required to achieve a fair and unbiased system.

Here are some specific strategies we use:

• Data cleaning: We make sure that the data we use to train our AI models is clean and free of bias. This means identifying and removing any data points that could contain discriminatory language or stereotypes. 
• Algorithmic fairness: Algorithmic fairness involves addressing bias in AI models. It’s essential to balance class representation in the dataset to prevent model favoritism. Techniques like resampling and stratified sampling help achieve this balance. Using fairness-aware algorithms during training and appropriate evaluation metrics is vital. Post-processing adjusts model outputs to ensure fairness by aligning predictions with predefined criteria.
• Human oversight: We also use human oversight to identify and correct bias in our AI systems. This means having humans review the decisions made by AI models and identify any potential biases.

8. Community Engagement

We want everyone involved. So we always aim to include impacted communities and stakeholders in the AI development process. Different perspectives make us better.

9. Responsibility of Policymakers

We care about policymakers, they are essential. We are keen to work together to create AI regulations that focus on ethics and fairness.

10. Continuous Learning and Ethical Training

We never stop learning. We stay updated on ethical guidelines and keep improving our ethical AI practices. 

11. Ethical AI Auditing and Monitoring

Our core team keeps things in check with regular audits. We make sure our AI solutions stay ethical.

12. Reporting Mechanism

If any community member spots something fishy, we’ve got a reporting mechanism. Which offers transparent and prompt handling of all reports.

Omdena’s Code of Ethics is our way of ensuring we build inclusive, democratic, and responsible AI solutions. We invite everyone to join us in creating a future where AI benefits all of humanity. Together, we can make AI a force for good.