From Ethics Zero to Hero !
How to Implement a Dynamic Process and Sleep Soundly at Night
The Struggle!
As an ethics & compliance manager, you may be feeling overwhelmed by the media hype around AI and wondering how you can make a greater impact on your business. You've probably noticed that tech giants are firing their best ethics teams, and you're worried about when AI technology will evolve beyond your control.
You've already taken some steps to address the issue, such as creating an ethics charter, an ethics committee, and training programs. But you know that these measures may not be enough to keep up with the rapidly evolving technology.
What can you do to have more impact in your company?
The answer lies in creating a better ecosystem that allows you to shine.
Here are some steps to get started:
Take an ongoing inventory of your AI applications in production, under development, or design.
Identify those with a strong ethical impact (e.g Bias, Explainability, Privacy)
Gather a diverse group of experts (e.g compliance, marketing, procurement, data scientists, etc)
Evaluate potential ethical “nightmares” that may come in from different perspective (users, customers, citizens, employee, other stakeholders)
Evaluate how these issues relate to the mission and ethical values of your company.
Categorize the results by theme (e.g.fairness, anti discrimination, transparency, accountability, respect, security, beneficence, Human in the loop, etc)
For each theme, evaluate the objectives to be achieved corresponding to different levels of learning (e.g Non-existent, Initial, Defined, Managed and largely used, Integrated and Proactively managed).
Identify at which level each AI application is located.
Evaluate the actions to be taken to move to the next level and define performance and completion indicators.
Periodically review results and progress.
Once this approach is developed and sufficiently robust in the scope concerned, evaluate the modalities to generalize it to the enterprise level.
📌 Some may question the point of this work, saying that data scientists know how to do their job and that it's too time-consuming. But don't let others influence your ethics and tell you what to do.
The key is to test and try new approaches, and to focus on the "how" rather than just the "what." Only by thinking about how to build momentum around ethics will you be able to create a more ethical business
As an ethics manager, you are the key to creating this dynamic. It's not something that can be learned overnight or imposed from the outside. It's a collective learning process that takes time, so start small today to be ready and armed tomorrow!
What do you think?
SAFE AI NOW
