Agnieszka Zimolag

Product Designer at BCG DV | Researching products for distributed intelligent ecosystems

Agile ethics. EthicsOps. Ethics OS. Minimal Ethical Product

Over the past few years, I have attended quite a number of conferences where ethics and technology were the main topics of discussion. In each of them, the concerns and dangers surrounding the ethics were mentioned but the thing that I was missing at the end was the actionable items. What type of set of actions should we be equipped with so that we can work towards more human-centric products and services? What I mean by human-centric is working in favor of the individual, societies, cultures and global population. So it is not only about the dark patterns of the UX. It is a layered structure that involves all parties taking part in the product/ business creation.
We should be looking at the entire business/ strategy if we want to talk about ethics.
(Design Ethics and the Limits of the Ethical Designer by Curt Arledge)
I have prepared a list of simple steps/ operations that could hopefully spark more practical conversations bringing new insights on how we approach building ethical solutions. I would like to also a question of what do we really mean by ethical and how it is different from the already existing domain of design/ product strategy?
“Our understanding of what ‘better product’ means has evolved. When we used to say ‘better products’ we meant those that were easy to use. Today, however, better products respect people and avoid doing harm.” Dan Brown — UX in the Age of Abusability
Digital ethics or information ethics in a broader sense deals with the impact of digital Information and Communication Technologies (ICT) on our societies and the environment at large. Digital Ethic by Rafael Capurro
Technology represents only the possibility of substance, so that we need the power from design to make it really become part of the field of life. Design is the process of trying to make the world friendlier to us clumsy humans; it is the effort to make the world more caring toward us, more accepting of us and so more morally acceptable to us. Ethics by Design, or the Ethos of Things by cameron tonkinwise
It seems that whenever we talk about the ethics we are simply referring to the design domain. However, design thinking needs to be implemented bottom-up, not only to the design cohort but to all the other decision-making parties included in the infrastructure and business layer. 
(Agile Ethics is an 8 step process, short-coded DATA-SCAPE https://humansinai.com/)
I would like to propose an agile form of implementing ethical solutions to your organization:
Define the meaning of the ethics for your customers and the core values of your organization.
Define the culture of ethics in your organization
Define the ethical threshold for the industries you want to work with and any third parties interacting with you
Define what could be unethical behavior in regards to your product/ business and define your plan to mitigate those risks
Include your customers into defining the ethics
Define what future you want to create for them
Ideate on the dark patterns
Find out how the ethical principles can be turned into your product/service principles by the individual cohorts:
  1.  Product
  2.  Engineering
  3.  Business
  4.  Organization
Define a roadmap and scope for implementing the ethical framework to your product
Set up processes that will ensure that with every iteration you are compliant to your ethical principles. Set up a team responsible for it
Set up a pipeline of the implementation of those principles ethical too such as making sure you do the proper data audit
Set up processes that will help you keep iterating on the principles and finding potential new risks that you have set up for your business/product
Ethics are also about perception. What message does your business convey? You need to have a strategy after setting up your EthicsOps for your ethics will be represented to the world, what experiences you want to give to people. Imply those principles into your branding as well as product design processes
Make a plan for when things can go wrong. Think about all the bad scenarios and make a plan how are you going to mitigate the risks
Don’t stop with yourself. Include other partners, service providers to comply with your standards. Bring the change to the industries around you
(https://digileaders.com/how-the-nhs-can-lead-the-world-in-ai-based-healthtech/responsibility/)

1. Ask questions about what future we want to envision.

Each product has the ability to influence someone’s reality. By the products/ services we produce we are creating a preferred form of reality for the humans. It is very hard to predict how eventually the product will change someone’s life but we can ask questions early on in the development cycle to mitigate certain risks and influence the outcomes of the usage. We can go one step ahead and imagine what future we want to create for the person using our service. 
How do we want to influence someone’s life? 
What values do we want to bring to someone’s life? What values do we want to introduce to society? 
Are we aware of the frictions this person is facing today? How our product is going to leverage those frictions? 
What type of community do we want to create?
What culture we definitely don’t want to create? 
And finally what is the roadmap of our product to achieve that?
A practical tool to use: Design For Meaningful Growth (PDF)
(Vijay Chakravarthy — Design For Meaningful Growth)
A practical tool to use: Ethical OS (PDF), Ethical Design Thinking (PDF), Tarot cards of Tech
(The Ethics Canvas by Wessel Reijers)

2. Create tools to include people in the conversation

If you want to create inclusive products you need to engage people with your products and include them in the creation of your new offerings. That becomes more challenging when creating B2B products, large scale platforms or institutional solutions. The impact of technologies should be brought onto an institutional level where the operations of governing tech should become institutionalized. The governing tech department should be busy studying the impact technology has on society and should research the tech patterns influencing people’s lives. It should create regulations, standards applicable across different industry sectors. They should be busy researching what changes could be introduced to save people from the dark patterns and make tech industry inclusive. It should create tools to bring people into the conversation so we have more decisive power about what future we want to envision for ourselves with the products/ services that we use every day. 
However, as Cennydd Bowles is saying in his article Thoughts on chief ethics officers : 
‘We need an exec’ tends to be a (slightly facile) default position whenever someone identifies a gap in tech company capabilities. But I think the best approach is rather more interwoven. A chief ethics officer would be too distanced from product and design orgs, where most ethical decisions are made; their duties would come into conflict with those of the CFO, who is already on the hook for financial ethics; and the seniority of the role would mean this person would be seen as an ethical arbiter, an oracle who passes ethical judgment.”
(Source: Cennydd)

3. Study ethics in-depth and make ethics inclusive as well

Ethics is a very complex field in itself. We have to understand the domain of ethics first in order to be ready to speak for the different populations. I think we should make the ethics inclusive to make sure that we cover different concepts of value, include different perceptions, culture and work around particular fields of application such as Bioethics, Business ethics, Machine ethics, Military ethics, Political ethics, Public sector ethics, Publication ethics, Relational Ethics, Animal ethics. On top of that, these ethical principles need to be studied in-depth to be adaptive to the specific industry/ product. In the end, these ethics need to be turned into programmable principles. Perhaps the field of Ethics needs a revisit too to be applicable to our vastly connected, distributed world. Each type of software solution should have its own ethical implications study in depth. Whether it is blockchain technology, AI/ML solutions, nanotechnology, data science, biotechnology each field should have its own ethical principles to which they are compliant and can be further used by the engineers working with these solutions. 
The framework focuses on the design process of the intentional approach:
  1. Define the problem being addressed and the desired outcomes
  2. Explicitly identify the ethical approach
  3. Assess the ecosystem of the desired outcome
  4. Determine the guiding design philosophy
  5. Determine if blockchain is an appropriate technology choice
Once the blockchain is selected as appropriate technology, the framework then moves iteratively through a detailed analysis of six root issues: governance, identity, verification/authentication, access, ownership of data, and security. At each stage, guiding questions serve to identify the effects of the design choices on the end-users and communities.
  1. How is governance created and maintained?
  2. How is identity established?
  3. How are inputs verified and transactions authenticated?
  4. How is access defined, granted, and executed?
  5. How is ownership of data defined, granted, and executed?
  6. How is security set up and ensured?
A practical tool to use: Designing Transparent AI (PDF), Ethics Tool Chest by MIT (PDF)

4. Specialize in making ethics programmable and easily implementable.

Apply your new principles into the engineering and design process. Think about a different type of people using your service. Think about different contexts they might be interacting with your service. They might have chosen your product over the others because of certain values/ reasons. They might be guided by different principles — respond to them by anticipating their needs. Anticipate people’s next moves to give them a variety of choices tailored to their current context and life situation. Create a checklist specifically for your product that could be part of your agile development process. Any time there is a new iteration being made the engineers/designers could go through this checklist to make sure they are following the guidelines you once created. The checklist should be constantly amended as your product matures and the service your offer change over time. As an example I will show how a data scientist could create his/her own checklist based on the Data Science Ethics Checklist generated with deon.:
A. Data Collection
A.1 Informed consent: If there are human subjects, have they given informed consent, where subjects affirmatively opt-in and have a clear understanding of the data uses to which they consent?
A.2 Collection bias: Have we considered sources of bias that could be introduced during data collection and survey design and taken steps to mitigate those?
A.3 Limit PII exposure: Have we considered ways to minimize exposure of personally identifiable information (PII) for example through anonymization or not collecting information that isn’t relevant for analysis?
B. Data Storage
C. Analysis
D. Modeling
E. Deployment
A practical tool to use: Ethical Design Thinking (PDF), Digital Data Governance Workbook (PDF)
People + AI Guidebook : https://pair.withgoogle.com/

5. Define the ethical rules of interaction based on your technology and the systems your product is interacting with. As an example, I am going to use distributed intelligent ecosystems.

(Agile Ethics for the Intelligent Ecosystems by Agnieszka Zimolag)
When interacting with other intelligent ecosystems we must comply with its ethical principles. Although agreeing to terms and conditions should be flexible. The terms should be bendable to allow for more interactions. However, there should be a certain threshold to remain authentic.
We as individual intelligent ecosystems we can change our ethical principles codebook based on the interactions with other intelligent systems. Nothing is set in stone. Experience should teach us something about the intelligent world out there. By doing that our identity becomes more adjusted to interact with more intelligent ecosystems.
With any slight change to our code of ethics, all the intelligent ecosystems should receive that update and comply with it in order to be able to interact with us again. 
Our intelligent ecosystem should possess its own standards about what types of intelligent ecosystems it wants to interact with.
The outcomes of the interactions between x>=2 intelligent ecosystems should be dependable and comply with the code of ethics of each system. As each intelligent system has its own values and favors one outcome of the other the consensus has to be reached over the preferable outcome
The ethics in distributed systems should be layered and that should also define to which principles the intelligent systems should comply to first and which principle can override another one. There should be local principles and global ones governing all. Every intelligent ecosystem should comply with global principles.
Ethically Aligned Design Conceptual Framework — From Principles to Practice

6. Build MEP (Minimal Ethical Product)

Before any product hits the market, founders usually decide on what the MVP looks like. Well, as AI-powered products are expected to turn into black swans within the next few decades, my assertion is that an MEP is very much needed instead of an MVP. Build a Minimum Ethical Product byCatalina Butnaru
Having MEP can become very useful just like defining MVP. It is going to help us to test if the set of the principles that we have established is going to be successful in the real world and what iterations are necessary going forward. MEP can be divided into sprints and with each sprint, a new set of principles required for our product and business embedded into a product can be prototyped and tested. 

7. Become innovator in your space

Just like products evolve and technology they are using, the field of ethics is going to change with it. The nature of each technology requires moral principles just like every feature that you design. The earlier in the process these guidelines are established the easier it will be to iterate on them later on. Make your ethical guidelines open source. In that way, the industry will more likely to adapt faster. You will become recognizable for compliance with the guidelines you set. And you will become the innovator in that field. You could set the future of the inclusion for the specific line of products, technologies, create collaborations and new opportunities as the programmable ethics will become a standard and those who are able to adapt faster will be able to set a new norm in the industry. 

Further reads:

If you are interested in learning more about my approach to designing for the distributed autonomous economy you can read my other articles titled The Instagram of Trust: How to Redesign the Architecture of Trust in Products published by hacker noon tech and Distributed self-sovereign identity published by uxdesign.cc.
Don’t hesitate to get in touch via email, or👇
🐦 Follow me on Twitter
Disclaimer: These are the projects that I do independently in my free time out of my passion and they are not connected to any organization and I am not collaborating with the companies mentioned in the article.

Tags

Comments

More by Agnieszka Zimolag

Topics of interest