In the AI chain … Whom to blame…
Salem Al Gharbi, Ph.D.
Transforming AI visions into a living reality | AI Experts | Leader | Strategic thinker | Visionary
Slowly -but surely- #AI #applications become part of our day-to-day activities. Starting from simple TV show recommendation systems to airspace and Gigafactory automation process. So, making a mistake is not unexpected … but who will be #held #accountable if this occurs??
For example, … if energy forecasting systems miss predicted energy production, leading to disturbances in the energy supply chain? Or The disease identification model gave false negatives putting patients in a dangerous health situation, … etc etc…
… whom to blame …? is it the AI model developer? What about the platform this model was developed or run under, it could be the cause for this fault? Did we think of the mathematical equations that represent the DNA of such AI algorithm? …?“the data is the AI fuel.” So, was the main source of the problems the data? If this is true, should the data provider, data quality improvement team be blamed … or the sensor manufacturer that did not collect valid data points?
From AI Ethical principles, this situation falls under…” Accountability & Responsibility”, which aims to hold ethical responsibility and be liable for AI decisions and actions that might result in negative effects. As a result, all parties involved in this process need to consider this during the AI project.
Until then … ensure that a human factor is involved side-to-side with any AI decision!??
Licensed Professional Engineer in Petroleum Engineering. Team-Player || Business-Savvy || Data-Driven || Value-Driven || Self-Driven
2 年When we talk about Artificial Intelligence and Human Intelligence, we separate them like one could replace the other. Right now, the concept of the augmented intelligence has been well accepted. Augmented intelligence?is a design pattern of people and artificial intelligence (AI) working together to enhance cognitive performance and decision-making.
Digital Transformation and Petroleum Engineering Consultant
2 年Interesting read Salem Al Gharbi. In any industry, there are several simple/complex tools and technologies used to manage the business, run and optimize operations, conduct predictions etc. These tools and technologies evolve with time and so do their utilizations. Every tool/technology has its own capabilities and limitations; thus, many industries do not use one single tool but multiple tools for automation, analysis, simulation and predictions to minimize uncertainties, risk and maximize value. The same concerns you have mentioned in the post are applicable to any old or new data-management tool, analytical tool, automation application, simulation solution and decision making process. The more confidence in a tool, the more expected utilization of that tool. AI is just a tool in a tool kit and it will always be. #AI #digitaltransformation #solutions #predictions #value #industry #data #management #application #simulation #technology #decisionmaking #tools
Management | Sales | Consulting | Artificial Intelligence | Automation | Engineering | Software | Digital Transformation
2 年This is a thought provoking article Dr. Salem. I think that the person that should be accountable and responsible is the person whom AI is serving. For example, if a plant manager trusts AI to autonomously run his plant (on his behalf) but the AI led to an explosion, then the responsible and accountable person should be the plant manager. If a CEO decides that there is no need for a plant manager and trusts AI to take this role, but the AI led to an explosion, then the responsible and accountable person should be the CEO. My two cents…
Driving Digital Transformation | Head of UR Data Management & Governance | Project Management | Proven track record in delivery | MBA Candidate | MS & BS in Software Engineering
2 年Great piece! Dr Salem Al Gharbi