top of page

The Ethics of Generative AI

This is a significant and subjective area and is posing a huge challenge to humankind. Every country and continent on the planet is now faced with the task of trying to apply moral and legal guardrails to the application of GenAI.  In the right hands it will be a force for good; in the wrong hands it will achieve the exact opposite. Now the technology is available there is an urgency to try and apply certain levels of control and legislation and as always (as we have seen with social media) legislation is often too late. It is therefore the responsibility of people like you and us to ensure we create and adhere to a common code of conduct on it's application. Google regularly update their AI Principles and it is good to see them leading the way. They have a list of seven clear principles they will adhere to, as well as naming things they won't use AI for. The summary is below. 

ethics about generative ai.jpg

  1.  Be socially beneficial

  2. Avoid creating or enforcing unfair bias

  3. Be built and tested for safety

  4. Be accountable to people

  5. Incorporate privacy design principles

  6. Uphold high standards of scientific excellence

  7. Be made available for uses that accord to these principles

ai robot in a library of a thousand law books.jpg

The things they won't apply AI to:

  1. Technologies that, or are likely to, cause harm

  2. Weapons or other technologies whose principle purpose is to harm humans

  3. Surveillance technologies that violate internationally accepted norms

  4. Technologies that contravene human rights

things artificial intelligence won't do.

Ethics, however, is multi-faceted and maybe the Google principles omit some other elements? As well as what AI is used for, we should also consider:​

  • How it is applied in line with strong ESG objectives

  • Who it is applied by, ensuring access to this technology has a thorough vetting procedure; and not be restricted to only those deemed wealthy enough to have access to it.

  • Where it is to be applied, leading to sovereignty and compliance of specific regions and industries.

​​​

It would be dangerous to wait for government legislation to catch up, as it may never actually get there due to the rapid pace that GenAI is developing. The responsibility, today, resides with the creators, deployers and users of this technology. There is a need for a code of conduct and self-policing, even if that resides inside of organisations to begin with. Ethics is the duty and responsibility of all of us.

The EU AI Act

the EU Ai Act.jpg

The European Union released their first iteration of their EU AI Act in March 2024. It goes some way to covering as many scenarios as possible and it is due to come into force in March 2026, a full two years after its release.  To our point earlier, in technology advancement terms, 2 years is a very long time, so there will need to be further iterations for it to remain current. That said, the EU did tackle this issue and produced something. The challenge is, it's 459 pages long, that's more than Harry Potter and the Prisoner of Azkaban. However, if you'd like to read it, you can do so below. 

bottom of page