2 Principles of responsible AI

It is essential to develop AI systems based on trustworthy, fair, and privacy-conscious principles. This video outlines six key principles: accountability, inclusiveness, reliability and safety, fairness, transparency, privacy and security. Discover how these principles foster trust and help create AI systems that respect individual rights and serve society.

Download this video clip.Video player: Video 2

Show transcript | Hide transcript

Transcript: Video 2

Voice-over:

AI is all around us.

It is there when you ask a voice assistant to play your favourite song, or when a chatbot helps you with a customer service inquiry.

It is guiding you home on your navigation app, and it is even working behind the scenes when your email filters out spam.

But have you ever wondered about the principles that guide the creation of these AI systems?

How do we ensure they are reliable, fair and respectful of our privacy?

The journey to responsible AI begins with trust, a trust that is built on six core principles: accountability, inclusiveness, reliability and safety, fairness, transparency, and privacy and security.

Accountability is not just about creating AI systems, it is about taking responsibility for their impact.

Clear roles and responsibilities are defined within development teams and across organisations.

Inclusiveness is about designing AI solutions with everyone in mind, ensuring that the benefits of AI are accessible to all.

Reliability and safety are achieved through rigorous testing, validation and continuous monitoring.

Safety measures include fail safes, error handling and protection against attacks.

Fairness is about treating all individuals equitably.

Regular assessments ensure that AI systems do not favour any particular group based on race, gender, or other characteristics.

Privacy and security are about respecting and protecting user data.Only necessary data is collected and users have control over their data.

Security measures are in place to safeguard against unauthorised access and breaches.

Numerous tools and resources are available to aid in the development, deployment and advancement of AI.

These resources and tools can guide developers through the responsible evolution of AI technologies.

They can assist with everything from identifying mistakes and assessing fairness, to exploring data and understanding the AI's decision making process.

The group of five gives a high five underneath a bar graph.

This ensures that the AI systems are reliable, fair and transparent.

One example that embodies these responsible AI practices is Microsoft Copilot.

It prioritises user data privacy and puts your data under your control.

It provides context aware suggestions, allowing users to understand how it arrived at a recommendation.

It aims to assist all users equally and undergoes rigorous testing to maintain reliability and safety.

The approach to developing AI is dynamic, not static.

Continuous learning from experiences, user feedback and advancements in the field are integral to improving AI systems and practices.

Collaboration and open dialogue with various stakeholders, including users, partners, policymakers and the broader public are needed.

Compliance with all relevant laws and regulations related to AI is a commitment, along with advocacy for thoughtful and informed AI policies that balance innovation with the protection of users.

Developing responsible AI is a process that places high importance on six key principles accountability, inclusiveness, reliability and safety, fairness, transparency, and privacy and security.

The goal of these principles is to create AI systems that not only benefit society, but also respect individual rights and values.

End transcript: Video 2

Download

Video 2

The journey to responsible AI begins with trust, a trust that is built on these six principles:

• Accountability: Define clear roles and responsibilities for AI impacts.

• Inclusiveness: Ensuring AI benefits everyone and is accessible to all.

• Reliability and safety: Extensive testing, validation, ongoing monitoring, and safety protocols with robust error handling.

• Fairness: Equitable treatment of all individuals, with regular assessments to prevent bias.

• Transparency: Enable users to comprehend AI decisions in order to build trust.

• Privacy & Security: Safeguard user data, collect only necessary information, and implement strong security protocols.

These principles aim to develop AI systems that benefit society while upholding individual rights and values.