There are times when I really want you, the reader, to know that I wrote something without AI. That the ideas and words came from my head to the keyboard to you. And in the reverse scenario, there are occasions when I want to trust that the author of messages I receive did indeed write it.

Here at Convictional, we call this idea "souls only". It's an environment or a communication artifact where the human communicators feel confident they are the originators of the content. No AI used.

In a world where AI tools are becoming integrated with all forms of communication, it's useful to pause for a second and think deeply about what we are gaining and what we are losing when AI starts taking responsibility for parts small or large within our communication. It's becoming increasingly challenging for us to discern whether an email or post we read is actually from the author's own hand and mind. We can no longer rely on the prevalence of em-dashes to tip us off. We don't want to be persuaded by or sold to by AI. I personally get disinterested when I read an email or post that has a hint of AI authorship. And in the reverse direction, it's increasingly hard for me to communicate to the people I'm writing for, that the ideas and words are my own. This is a big problem. The status quo here isn't great for society. And this same problem will soon come to video, with better and cheaper AI deep fakes. This will put a value on knowing provenance and authorship.

I understand that there's a tradeoff being made, namely: the writer trades trust and authenticity of authorship for AI's speed and "quality". I put quality in quotes here intentionally because I believe writing quality will shift over time to mean something different than it does today. AI is a tremendous tool for brainstorming and there is value in collaborating with AI to find that right word to use or to help articulate a feeling that one is having a hard time expressing. AI is a great dictionary, thesaurus, spell checker and editor. But it can also "put words in your mouth." It is increasingly hard to know how much AI was used and how. And that question creates doubt in the reader. And that doubt erodes trust. And trust always was and will continue to be the currency of commerce, and relationships.

There are contexts where souls only mode is appropriate, when the authentic human judgment of the writer is essential, or said another way, when the judgment of the writer is the entire point of the communication. Examples include communication of company decisions, a teammate's review of your work, and the expression of emotion or feelings to someone else. In the sales or marketing context, communication of your unique value to customers as well as the reasons why customers should trust you or your organization are more meaningful if actually authored by you.

The closer a communication is to the author's unique, personal judgment, an argument can be made that the author and reader should expect less AI usage. We call the summation of these judgments within an organization its ethosystem. In the future, we think most knowledge work within companies will be within the realm of the company's ethosystem. Decisions, judgments, opinions, debates on quality, values statements, criteria ranking, etc. Employees' time during the work week will increasingly be spent on this type of cognitive work, and AI tools will take on the non-judgment, executional work. And it is your company's ethosystem where human authorship matters most.

Souls Only Technology

To avoid the most obvious case of hypocrisy, as well as showing how useful something like this may be, I rigged up my web cam to record my writing of this essay and used screen capture software to record the entire writing process, including my screen and hands typing. This is an absurd and uncomfortable way to author this essay, but I am gaining something very valuable in doing this. In creating an artifact that can be inspected by the reader that shows I did in fact type out this essay, I hope to indicate that these ideas and words came from my soul to yours.

Whatever technology eventually gets developed for human authorship will need to address a few things beyond just provenance. Ease of use, ease of verification, and comfort levels around any surveillance used will need to be addressed. There will also be an "arms race" in the pursuit of showing human authorship with increasing AI capabilities. ("Maybe Bill had AI generate the video of him typing this essay?"). Ultimately, I don't think we'll ever get to a 100% accurate proof of human authorship or at least won't desire the level of surveillance required to get there, but I'm certain technologies and norms will emerge for the purpose of building trust between humans.

When the use of AI is appropriate and desired, there will also be ways for the writer to communicate, and the reader to understand exactly how AI was used. It will be useful to see where human judgment is separate from the AI-assisted articulation or presentation of that judgment. Norms will develop that indicate what types of AI use, thesaurus and subtractive editor perhaps, are acceptable without eroding trust.

We believe that there needs to be new innovations in the area of communication provenance. At Convictional, we are building ways to help authors communicate to their readers that the text was written by them, and not Claude, Gemini or ChatGPT. We're giving ways for people to opt in to a souls only mode when appropriate. The same way there is value in struggling through decisions, there is value for both the author and reader when writers "struggle" with their own ideas and own words.