As a small team of volunteers, we need to leverage automation and AI as we can to improve the ability for those who are using the site to find relevant candidates for open positions to best be able to find them and quickly review their summary information.
We are looking for those who know how to leverage AI to explore the tools available within and outside of this Discourse platform to perform tasks such as:
Writing summary paragraphs of the replies on a nominee’s profile and pinning them as the first reply to the nomination profile
Calculating the number of characters in a nominee’s profile (first “topic” post) to assign a tag with this number to represent the level of effort and degree of completeness of each nomination profile
Allowing for people to sort nominations by these number of characters knowing that they will learn more from those with more complete nomination profiles
Exporting data out of Nominees for the People
Automatically creating tags based on key aspects of the profile (such as agency listings), to improve search ability and decrease the need for volunteers to manually tag profiles
Other ideas - we are open
We’ve reached out to Discourse for suggestions on many of the above topics and have a list of written resources they have pointed us to, for our AI Guides to use as initial resources on how to summarize and more. We do have one volunteer administrator who is keen on learning how to use AI better, and the AI Guides that come to us will not only help solve these problems but teach this administrator how to do it as well.
To express your interest and share your skills and ability to help enable the following AI tasks, please go to your Inbox on this site and message Phil at PCunningham.
Another idea to add on would be a way to show positions that are available and potentially remove(well not remove, but just “grey out”) those other candidates wanting those positions that are filled. That way you dont keep having people talking about the same nominees that arent selected. Example: Kash as FBI Director, now any other nominees on this site for FBI director would be “locked” or “greyed out” so no one can comment or vote for them.
Yes, if this project were funded it would be a great way to deploy some staffing resources to do that work. Right now we’re a team of part time volunteers. Let me know if you or anyone you know would like to help make that happen.
OK, so you want Artificial Intelligence programmed to think a certain way, maybe in support of big pharma or industrial farming, to literally steer people to candidates it “thinks” they would like?
If that isnt propaganda and manipulation I don’t know what is. Search the term ATF to vote for ATF director. Done. No AI propping up people that paid money to be seen at the top of the list.
That, is, literally, the, point, of, this, site.
Asking AMERICAN CITIZENS WHAT THEY WANT, WITH ZEEEEEEERRROOOOO INFLUENCE!
Seriously though, are you an AI trying to influence human behavior? Asking for a friend.
I’m the CEO of generative AI company. While we provide proposal automation for federal government contractors, our technology could easily be adapted to assist with your task. You can email me at [email protected] , and I’d be happy to assist. - Arthur
I share your healthy skepticism toward AI and no, I can assure you that that’s not what we have in mind at all: we just want to use AI to consolidate duplicate nominations for the same person or summarize replies, for example (some people have hundreds or thousands!) On this site, our role is not to select people or make recommendations to those who do the evaluating and vetting on the transition team. Rather, this site is a supplemental resource to any other information the transition teams have as they decide who they want to further vet or interview.
So, no, not using the AI to influence human behavior… just make tasks humans could take lots of time to do go faster. Only after efficiency gains.
I don’t recommend using AI in such a way as it will often let gems slip through the cracks if used in such a manner; however, that does not mean that AI shouldn’t be utilized.
Simply copy the applicant’s information into the AI and ask it to summarize it into a single paragraph.
I don’t recommend assigning a tag of “completeness to the post,” however, utilize a database of known keywords, things that are typically found across applications. Then have the AI detect the amount of words around such keywords and tabulate this number. That is the best way to determine original content from doctored content. The reason is that many applicants would rather just use keywords to highlight their experience but that alone does not give the application any value. You could refer to this as an “Intent Value” as the AI will be estimating the amount of words apart from keywords.
Use the “Intent Value” from all the candidates and average them out, candidates sitting within the 50th percentile would be candidates who overutilized keywords and failed to include any original content. Candidates closer to the 75th percentile should be the candidates who utilized keywords but expanded upon the meaning. By contrast, candidates in the 25th percentile should have used fewer keywords than average and have had more original content. This can be referred to as the “Intent Value.” I’d like to emphasize that such a process would not immediately give candidates credence to their credentials; it would only show who actively spent more time creating unique content compared to the rest of the population.
Utilize the summarized assessment of each candidate, take the information gained through the application, and save it in a database based on the “Intent Value” in a more digestible system. If necessary, create a master file based on the “Intent Value” in Google Docs and then create a hyperlink of each candidate’s information and place it accordingly. I’m assuming you’d already have something linking this on the back end of the website as you’d already actively storing candidates in separate web pages.
Refer back to the keywords, you should utilize the keywords as a means for categorizing individuals based on their perceived talents so that you could later create a list based on the agencies they might have the qualifications for. We can call it “Keyword Value.” Create multiple Google Docs, one for each agency, then place each candidate according to their “Keyword Value” in the proper destination(s) according to the “Intent Value.” In such a scenario, you will be able to maintain an accurate record of both “Keyword Value” and “Intent Value” without the two overlapping and negatively interfering with one another too much.
…here am i, at your service. I am a efficiency expert, basically a ai novice but i seem to have a real nack for it. Essentially the reason i landed here in the first place, me strumbling into leveraging ai, descoving i am a natural promt engineer/prompter.
Thanks @Beau-Garrett - will reach out to you if we have a need. Right now the AI tool within Discourse is able to summarize comments well, and we’ll see what else we may need. Thank you.
This is a good idea. It would be helpful to color code all nominees. Chosen nominees can be highlighted in green while others (who are out of the race) can be highlighted in grey. People who are still in the running can be left as they are.
I don't know if I agree about the voting thing though (maybe partially.) Those who have been appointed should be removed, but people who aren't nominated for 1 specific position should be able to collect as many votes as possible (since they still have the ability to do other jobs.) People should be able to cast votes until all of the role are filled. Removing this option wouldn't be fair or democratic. It would have to be done with caution so people still have opportunities. This may not change the appointments, but it would still highlight individuals involved in the process (without cutting them out.)
…here am i, at your service. I am a efficiency expert, basically a ai novice but i seem to have a real nack for it. Essentially the reason i landed here in the first place, me strumbling into leveraging ai, descoving i am a natural promt engineer/prompter.