This is from an episode of the “Everything AI and Law Podcast” where the host interviews Dr. Michael Akinwumi, an expert in the AI sector and the Chief Responsible AI Officer at the National Fair Housing Alliance. The discussion mainly revolves around the concepts of Responsible AI.
Key speakers
- Tolulope Awoyomi (Tolu): Lawyer & Everything AI and law Podcast Host
- Dr. Michael Akinwumi: Chief Responsible AI Officer at the National Fair Housing Alliance
Takeaways
Takeaway 1: AI systems are inherently socio-technical systems influenced by human decisions.
Dr. Michael Akinwumi emphasized the human aspect of AI, refuting the notion that AI systems are “black boxes” beyond human comprehension. He explained that AI systems are socio-technical systems, indicating their intricate connection with society and the fact that they don’t operate in a vacuum.
“There are different components right but I think it’s about four or five components. So to say, one you need data… two you need your problem statement… three you need code… four you need compute resources… five you need a policy that is going to implement the decisions coming out of the AI system,” Akinwumi explained. He adds that, “It does not really mean that everybody has to be a coder or a developer… because AI technology is going to disrupt everything, it’s impacting everything.”
Takeaway 2: The principles of Responsible AI include non-discrimination, privacy preservation, reliability, explainability, and governance.
Dr. Akinwumi explained the five principles that make up Responsible AI in detail. He stated that AI systems in housing and landing must ensure non-discrimination, preserve individual privacy, be reliable, explainable, and have a process in place to govern the system.
“By responsible AI, we’re talking about fairness or non-discrimination…you also need other pillars or principles. The second principle is privacy preservation…the third area for us is what I called reliability… the fourth principle is the principle of explainability… and the last principle is the principle of governance or contestability,” said Dr. Akinwumi.
Takeaway 3: Mathematics is a foundational aspect of AI systems.
Dr. Akinwumi stressed the role of mathematics in AI and coding, explaining how it forms the connection between data and the problem being solved. He stated that at the heart of AI systems, mathematical equations are used to identify patterns.
“Between data and the problem statement where you start peeling the onions and then you realize that the core of it or the heart of it is the model right…the heart of some of these generative AI platforms…you will see that it is really a set of mathematical equations,” he noted.
Insights surfaced
- Responsible AI has five principles: non-discrimination, privacy preservation, reliability, explainability, and governance.
- AI systems are not black boxes. They can be understood and explained, depending on the level of expertise and specialization.
- Building AI systems involves five components: data, problem statement, code, compute resources, and a policy for implementation.
- Mathematics plays a crucial role in AI, particularly in coding and problem-solving.
- AI systems are socio-technical systems that operate within society and have social impacts.
- The future of work is heavily influenced by AI, and it’s important to prepare for this shift.
Key quotes
“By responsible AI, we’re talking about fairness or non-discrimination one and then two privacy preservation three uh explainability four uh reliability that is safe and also valid across all groups and five um governance.”
“There’s no mystery [in AI systems] it just depends on the level of expertise, the level like the specialization and all that and also being able to also ask the right questions right so that people can dig and find answers based on where or areas are the.”
“The code is a representation of the logic that is represented as math right so in between is where you need the code so that you can make that connection to the problem statements you can evaluate how it is solving the problem and all that right.”
“AI system is a social technical system it is a system that operate it does not operate in vacuum right so it is connected to the society so it has social impact.”
“Technology is a disruptive force that is going to uh you know touch everything and if you’re thinking about you know where what impact we can create right you need to think about what areas you want to focus on.”
“Hard work does not kill but it hardens you and prepares you for a better tomorrow.”
You can catch up on the podcast episode here.
Comentarios