Since 2018, I have been engaged in AI-related work, and have a cautious but stable attitude towards the development of AI, and are also very interested in unstructured trends. Now, initial concerns about robotic takeover have faded, replaced by discussions around the ethical issues that bring about incorporation of artificial intelligence into everyday business structures.

A whole new set of roles are needed to deal with ethics, governance and compliance issues, all of which will have great value and importance to the organization.

Perhaps the most important of these are AI ethics experts, who need to ensure that AI systems meet ethical standards such as fairness and transparency. This role will involve the use of dedicated tools and frameworks to effectively address ethical issues and avoid potential legal or reputational risks. Human supervision to ensure transparency and responsible ethics are crucial to maintaining a delicate balance between data-driven decision-making, intelligence, and intuition.

Additionally, positions such as AI workflow designer, AI interaction and integration designer will ensure AI is seamlessly integrated into the ecosystem, with transparency, ethical considerations, and adaptability prioritized. An AI supervisor is also needed to monitor the agents and arbitrators of the entire stack, i.e. the decision-making elements of AI.

For anyone who set out to incorporate artificial intelligence into their organizations and wants to ensure the responsibly introduced and maintained the technology, I recommend referring to the principles of the United Nations. These 10 principles were formulated by the United Nations in 2022 to address the ethical challenges brought about by the growing popularity of artificial intelligence.

So what are these ten principles? How do we use them as frameworks?

First, don't cause harm

Adapting to technologies with autonomous elements, the first principle focuses on deploying artificial intelligence systems in a way that avoids any negative impact on the social, cultural, economic, natural or political environment. The design of the AI ​​life cycle should respect and protect human rights and freedoms. The system should be monitored to ensure that this condition is maintained and that there is no long-term damage.

Avoid using AI for the sake of AI

Ensure that artificial intelligence is used reasonably, appropriately and not overdoing it. It is easy for people to over-enjoy the use of this exciting technology, but they must balance AI with human needs and goals, and must not be at the expense of human dignity.

Safety and security

Security risks should be identified, resolved and mitigated

Continuously implemented throughout the life cycle of artificial intelligence systems. AI should adopt a sound health and safety framework that is exactly the same as other areas of the business.

equality

Similarly, the use of artificial intelligence should also aim to ensure equal and impartial distribution of benefits, risks and costs, and to prevent any form of bias, deception, discrimination and stigma.

Sustainable Development

Artificial intelligence should be committed to promoting environmental, economic and social sustainability. Its negative impacts should be continuously evaluated, including its impact on offspring.

Data privacy, data protection and data governance

An adequate data protection framework and data governance mechanism should be established or strengthened to ensure that individuals’ privacy and rights comply with legal guidelines on data integrity and personal data protection. No artificial intelligence system should infringe on the privacy of others.

Human supervision

Human supervision should be guaranteed to ensure that the results of using artificial intelligence are fair and just. Human-centered design practices should be adopted and humans should be empowered to intervene at any stage, decide how and when to use AI, and overturn any decisions made by AI. Although the UN's proposal is dramatic, it is completely reasonable that no decision about life and death should be left to artificial intelligence.

Transparency and interpretability

In my opinion, this is part of the code of equality. Everyone who uses artificial intelligence should fully understand the system they use, the decision-making process of the system use and its impact. When AI makes decisions about its rights, freedoms, or benefits, the individual should be informed and, most importantly, explained in an easy-to-understand way.

Responsibility and obligations

This is the whistleblower principle, covering audits and due diligence and protection of whistleblowers to ensure that someone is responsible for the decisions and uses of AI. Governance should be centered around the moral and legal liability of humans for any AI-based decision-making. Any decision that causes harm should be investigated and action should be taken.

Inclusive and participatory

Just like any other business field, an inclusive, interdisciplinary and participatory approach should be adopted when designing, deploying and using AI systems, including gender equality. Stakeholders and any affected communities should be informed and consulted and informed of any benefits and potential risks.

Building your AI integration around these core pillars should help you feel assured that the first step you get into AI integration is to build on ethical and solid foundations.