David Armano

Just as learning platforms, tools and staying on top of AI topics is a dynamic process—so are the discussions and decisions made around ethics. Some, such as bias, privacy and security are foundational and unwavering—but other areas will evolve along with the industry, such as celebrities working with AI entities, platforms and providers to license their likeness. An AI Ethics framework should be considered "living" documentation, meaning it is updated and revisited frequently

It's important to use consistent language when defining a measurement framework which is a helpful tool in measuring the impact of anything for example: metrics are the things we keep track of at the most granular levels (views, impressions, shares etc.) KPIs: Should be a little higher level i.e. "raise awareness", "efficient use of dollars", "generate leads" Business Impact: should be the ultimate goal of what is being measured, ie revenue earned via sales or money saved via efficiencies Tying metrics, KPIs and business impact together is key and this is where things like attribution come in as well as having data tracked at all points along the customer journey

Garbage in = garbage out as the adage goes. This means that the data we input into things like machine learning systems not only should be ethical and as free from bias as possible, but it needs to prioritize quality over quantity. A higher quality data set can be more valuable than a data set which is five times the size, if the data is less relevant, inaccurate or gathered using flawed methodologies

Every organization needs to have an updated code of conduct and ethics to reflect the new realities that are constantly emerging around A.I. In fact, policies around ethics should be considered “living documents” meaning that they are updated as frequently as the technology. This is one of the most important factors of an organization having “humans in the loop” and is perhaps the single place where automation or A.I. involvement should not be applied. It is the checks and balances system for the organization. Likewise, every member of the org. should have access and be encouraged and proactively updated as the ethics are revised and revisited.

With Open AI’s latest multimodal functionality, you can now start your content with a visual such as an info graphic and ask GPT4 to write an article for that graphic. You can use additional prompts to expand on areas of the visual but overall, it will analyze the visual and create the content without you having to describe the visual in detail. Currently this only works for subscribers of the paid version

When using LLMs to generate an article, it’s important to establish “voice” and style early on. You can train your LLM on your own voice if you are published and your content can be found. You can also have it write in the voice and style of others whose published works are accessible. There are ethical considerations here of course. For example if it’s your own voice or you are ghost writing for an author, it’s straightforward. If you are emulating another voice that is not your own, it is less clear ethically.