AI As Co- Collaberator

Something that is interesting to note is that when a properly structured research project starts, it usually has a defined scope and a presented goal.However many honest research projects do not end up with their presented goal. Their findings may lead them in other directions. Does this mean the research project failed? No. What you have here is the observation that the stated goal provides a framework for discovery.And judging the discovery against the original goal helps you to define what it is you actually discovered. Otherwise you have discovered a formless void. Congratulations you have made a discovery but you cannot define it.

And so in this, working with AI follows the same trend, if you will. Meaning that if you don’t have a scope, if you don’t have a goal, if you don’t have clearly defined responsibility domains; your results will be murky or blurred*. By responsibility I do not mean that the AI takes responsibility for being wrong; that’s not what I mean. What I mean is that when the AI occupies a specific domain or role within the workspace, then their responsibility is very clear. It narrows their focus to that particular domain.Which also makes defining the human role in the work space all that much more prescient. This describes the difference in utilizing an AI to write an extremely technical email while giving it a very well-defined prompt and then just letting it write the email as opposed to writing a theoretical research paper or, in this other example from paulolden1 , where you can write a novel. All these are valid examples and are showing, not really empirically but by positive examples, how you can do it. I’m not going to say empirically because we’re not actually testing things. Well I am. I’m running my own workforce in the background and researching how everything plays out.And I plan to post some of that documentation eventually but that’s one of the reasons why I asked the question of what the community thinks of AI collaboration.

But yes defining the human’s role in the field means that you still have to define the difference between one prompt, one output and a long series of interactive prompts and outputs. This gets a lot more complex when you look at paulolden1’s example of utilizing multiple AIs in a single project.

i make sure that i read everything that is produced from the sessions i have with AI before i finilize documents. i do not have a ‘instruct then print’ mentality. mainly because at the end of the day, what ever they produce, i am responceible for it.

which is one reason i utilize AI in 2 fieilds. Program Development, and Theoretical Research from Argumentation. program development/building is self explainitory. but my theoretical research rarely starts from someone elses observations. if its my observations, my arguments, and my assertations, then im only responceible for what i think. that limits the posibility that the AI will ad hoc - copy and paste someone elses argument in to my work. which can happen.

so yes, complacency is a big deal.

basically, Humans provide Vison, Direction, and Scope Management in most AI Roles.