AI products become popular on the basis of the trust of people in these systems. People and professionals use AI systems to ease their lives, while they can rely on these systems to organize work and business. AI systems can be used for profession, news, and entertainment, while some rely on AI tools to provide them with recommendations of products, reviews, and composing emails, to mention a few.

What is trust in an AI Product?

Trust in an AI product can be described as the continued robust service to clients who were given the right to use AI products and even in times of uncertainty and chaos the product delivered the results correctly, appropriately, and fairly. The right to use can be obtained on the basis of some agreement or some monetary fees that were paid.

How much trust do people have in AI systems and how fair are AI systems people use?

Trust is one thing and fairness is another concept and both cannot be mixed. A trusted AI system is often fair however a fair system may not be a publicly trusted system. However, a combination of both trust and fairness in an AI system is the best combination.

How is trust built on an AI product?

The users use the AI systems over time, find it useful, spread word of mouth, others use it, measure its price versus usefulness and test in good and difficult times. This is how the audience grows with time with different kinds of marketing strategies.

Let’s take an example of the physics of trust!

When we trust someone we give a part of ourselves or some of our belongings, or our finances in the hands of the other actor. The actor can be a human or an organization. The decisions made by an actor affect us in many ways. This can lead to beneficial impact or may lead to loss. Another factor is the performance of the trusted system, which builds more trust once the performance is sufficient and above.

How trust affect the AI systems?

People use the systems they trust. There are cultural variations in the usage of the same product in different places which may affect AI product’s usability as well as profitability.

What are the ethical principles in trusted AI systems?

Here are some ethical issues illustrated in Bartneck [1].

• Systems should be beneficial and do good.
• Systems should have autonomy for people. That is AI systems are here to help humans decide things for themselves. Machines and algorithms can suggest but should not make decisions for human on their own unless asked for.
• AI should act in in just and unbiased way.
• Ethical issues such as abortion, and capital punishment, have their own controversies. These moral decisions are difficult to make even for the best AI systems. However, if the difficulty level is reduced then moral decision made by AI can be noted and shared for performance.
• Hate speech should not be part of any AI product.
• The system should be explainable. That is the output should be transparent and reproducible under the same conditions, and reasoning could be provided.
• Accountability of the processes followed by an automated or AI system.

Examples of such AI systems include creditworthiness determination using AI, determining high-risk patients, and use in courts to prioritize cases to be sent to a judge, recommender systems, airplanes, and so on.

Key area of AI systems to become popular and be used over a long time is Trust and Fairness in these systems. Companies make their products with this insight to be used by as many users as possible and at affordable prices. This indeed makes companies profit and gives them a name, and fame. Once at this level, such organizations can make products available and affordable to spread to more masses. People shall use it for the trust and fairness that so many others have in this system.

Note 1:

This is a review with our own comments and recommendations. The insights are based on a chapter from the latest book on AI Ethics, by C. Bartneck et al. (2021) and has been self-written.

Note 2:

The article in duplicate is present on the author’s medium account as well.

References

C. Bartneck et al., An Introduction to Ethics in Robotics and AI, SpringerBriefs in Ethics, 2021, https://doi.org/10.1007/978-3-030-51110-4_2