The Idea in 60 Seconds

  • The first thing companies do to enable them to deliver AI projects is develop an ethical framework. 
  • I produced ours for the team I work with.
  • Queensland Government has a set of 6 ethical pillars that they ask agencies working with AI to oblige : Alignment with the public good, accountability, transparency, security, privacy.
  • Most commercial teams in the private sector seem to adopt an almost identical set of ethical considerations.
  • IMHO, fairness is the ethical pillar that matters most – by a long way. 
  • Fairness is the foundation upon which all other principles—public good, privacy, security, accountability, and transparency—are built.
  • Fairness is not just a principle; it’s a decision-making criterion, derived from our own, internal sense of right and wrong. 
  • If we act fairly, we naturally align with the public good, respect privacy, ensure security, and act with accountability and transparency.
  • The other pillars are important, but only as checks and balances to ensure fairness is applied consistently and appropriately. 
  • Fairness, is the core decision making criteria. It’s the compass that guides every ethical decision.

Fairness: The Core of Ethical Decision-Making

I wrote the AI Ethics framework for the organisation I work for. That involved research on how other government entities had approached the ethics of AI.

My goal, when I wrote the document was not just to consider the ethics of AI but, pragmatism. I wanted to provide a practical framework which would allow us to start delivering AI solutions as quickly as possible. i.e. my goal was to get going, not philosophical debate.

That said, of course, I had my own views on wthe subject matter. In my opinion, of all the ethical pillars that have been suggested, one stands out as the most important. Fairness. At its heart, in my opinion, fairness is about balance. It’s about weighing competing interests, allocating resources equitably, and making decisions that are justifiable to all stakeholders. Importantly, fairness is also the only deeply personal consideration. It’s an internal compass that guides our sense of what is right and wrong. In short, we each, individually know what’s right and what’s wrong. It is that that should inform the decisions we make, across the other pillars.

Considering the other pillars:

  • The Public Good: 
    Working for government, we should not even consider activities which are not in the public good. The only ethical hurdle here is to ensure we are aware of and acting in line with the most fundamental premis of our jobs. 
  • Privacy: 
    To me, Privacy is a legal constraint, not an ethical one.
  • Security: 
    Security requirements are not unique to AI projects, they are required for any IT project. I don’t consider Security inherently an ethical question.
  • Accountability: 
    We are automatically accountable for any AI work we do. Taking accountability is not an ethical demand. It’s the natural result of us building an AI solution.
  • Transparency: 
    The level of transparency we need to put in place is often dependent on the situation.

I think it’s self evident that fairness is not just one pillar among many—it is the foundation, internal, ethical basis upon which the entire ethical framework rests.

Why Fairness Matters Most

Consider this: if a decision is fair, it will naturally align with the public good. It will respect privacy, ensure security, and those involved will be honest about what they’ve built (accountable) and share the ‘right’ level of information with customers ad staff to make it transparent. 

Fairness is the principle that ties everything together. It’s the glue that holds the ethical framework in place.

I Do Agree the Other Pillars Still Matter – As Checks And Balances – But They’re Secondary

While fairness is the core principle, the other pillars play a role as checks and balances. They ensure that fairness is applied consistently and appropriately, and they help us avoid blind spots or unintended consequences. 

Conclusion: Fairness as the Compass, the Pillars as the Map

In my view, fairness is an ethical compass that guides every decision we make when we’re building AI solutions. It is the principle that ties all others together, ensuring that our actions are justifiable, equitable, and aligned with the public good. If we don’t have an internal sense of right or wrong, we could compromise on any of the other pillars without discomfort.

However, I do agree the other pillars—public good, privacy, security, accountability, and transparency—have value. They act as checks and balances, ensuring that fairness is applied consistently and appropriately. They remind us to consider the broader impact of our actions, respect individual rights, and maintain public trust.

In the end, in my view, fairness is the core principle, but it does not stand alone.