AI Utility

As someone who has created marketing content in agency settings, with over 30 clients, I see the value in using language models to speed up content creation. Identifying the limitations of a tool, from my perspective, doesn’t rule out good uses. ChatGPT is great at finding certain types of issues in code and improving resumes for job-seekers.

I think one of the best use cases for ChaPT is marketing copy, so long as the prompters are great fact-checkers and are dedicated to editing the content rather than just copying outputs. Unfortunately, we know human nature is lazy and we tend to over-rely on tools far too often.

Even this use case can hardly address the IP issues related to training these types of models on text real humans created with no respect to their creativity. I suspect even my past content for social clients might have been part of the training set for ChatGPT, but unfortunately, we may never know. ChatGPT hasn’t disclosed what type of data was used to train it, missing the AI Principle of transparency by a mile.

The Role of Ethics

I think since AI has such a high utility, we have to move beyond “strongly considering” ethics to fully embedding ethics in the AI design and development lifecycle.

Companies are rewarded for using AI tools in harmful ways such as contracting with governments and law enforcement agencies to continue a cycle of over-policing marginalized (often Black and brown) communities and immigrant populations. In some cases, these automated tools work really well, at reinforcing the status quo (which are oppressive systems).

In other cases, these tools have performance issues and just don’t work as well for some groups of people, but when the context is surveillance capitalism, we don’t really want these tools to “accurately” discriminate against us.

Should we be building AGI

One of the biggest questions we have to answer at this crucial point is if building AGI is our goal? Forget about the Terminator-type sci-fi examples, but what does AGI even look like?

The best answer comes from a pioneer in AI Ethics, Timnit Gebru. Spend the time watching her talk that covers the history of AGI efforts and their deep ties with eugenics.

Plans for this Newsletter

Just so you know, there are a few types of things I plan on covering with this newsletter including:

• Responsible AI 101

• Long-form reads

• Really Cool Work from others in the space

• Courses and Learning materials

• Ways to get involved

_{That leads me to wonder…}