The End of Credentials
AI Sees the Person Behind the Paper
Our ineptitude in getting at the record is largely caused by the artificiality of systems of indexing. When data of any sort are placed in storage, they are filed alphabetically or numerically, and information is found (when it is) by tracing it down from subclass to subclass... The human mind does not work that way. It operates by association. - Roy Ascott
Labels set our realities. We vote for people because of the party label attached to their names, often with no understanding of the person or interests behind that label. We choose our employees the same way: based on the labels attached to their resumes. Food, clothes, cars, are all marketed to us through labeling shorthand.
A lot of the superficial criticisms that are leveled at AI revolve around its ability to generate cheap content that mislabels its origins. However, it is precisely AI that holds the key to seeing beyond incomplete or even deceptive thinking. It makes it easy to see beyond labels.
Puncturing the Label
For example, I was offered an “opportunity” to be a “strategic advisor” for a tech firm on LinkedIn. I did not know the person making the offer and I’d never heard of the company. It was an enticing offer. It looked legitimate. It didn’t seem to hurt to at least follow up.
Instead, I gave both Gemini and Claude the text of the offer and asked them to investigate the person and the company. They found no record of either. The address provided was an empty office space. The labels looked good, but the reality looked different.
Another kind of example comes from a report on PBS about “declining education scores.” In it, they interviewed Harvard professor Thomas Kane, who argued that the undermining of the testing regime coupled with social media (cellphones particularly) have undermined students’ capacity to learn.
I was skeptical. There are a lot of studies that have found little or no connection between technology use and educational performance. I also see the products of the testing regimes imposed in high school in my classroom. They are certified as being ready to learn complex issues, but they are, by and large, not good thinkers.
That doesn’t mean they are dumb. Far from it. It’s just that no one has ever pushed them up the ladder of Bloom’s taxonomy.
Large language models do those lower levels well. That means we are certifying graduates who are likely to be replaced by AI. If AI can do it, why should I pay for an expensive employee?
I also don’t need school for this level of thinking myself. When I watched the PBS segment, I knew something was missing, but I didn’t immediately know what. So, I went to Gemini and had an hour-long conversation about the professor, his findings (which aren’t wrong), and the missing pieces.
As a social scientist, I knew that there were probably additional systemic factors at work here and, sure enough, funding for early childhood and support programs for children in school dropped precipitously during this period, victims of state spending cuts in the wake of the Great Recession of 2008-09.
The problem with testing without context is that we blame the teachers for the scores, not the legislators who pulled the rug out from under them. The labelling system doesn’t tell you who is responsible for the food poisoning, so you blame the restaurant, not the butcher who supplied the bad meat.
The Classroom Reality Check
But there are bigger problems. AI requires students to operate at the evaluate, analyze, and create levels of Bloom’s taxonomy. As my course this last semester demonstrated (small sample size alert), when I switched their cognitive tasks to those levels, my students really struggled.
How do I know this? AI can take me behind the labels of their grades and show me context. It helped me design assessments that precisely measured certain skills over the course of the semester. Then it helped me analyze the student outcomes on those assessments.
These were not simple, “did they get the answer right on the test” kind of assessments. I wanted to see beyond that label to what was really going on in their heads.
When I asked the students to start assembling their final portfolios (a website) they were suddenly forced to analyze and create (the top levels of Bloom). They seemed to be doing fine up until then but when I flipped that invisible switch, they suddenly weren’t.
I have personally always been good at evaluating and connecting. However, this has also been my biggest challenge as a teacher to understand my students’ mindsets. The labels I used before didn’t help me. AI let me see the patterns that I struggled to see before.
I was able to put all their results anonymously into an LLM and generate contextual analyses. If a student passed Module 3, where the analyze and create layers kicked in, they had a 100% chance of passing the class. If they didn’t, that number went down to 14%.
The SAMR Threshold
Coincidentally, I see the same pattern when it comes to AI adoption among my colleagues and the world at large. My friend, Ruben Puentedura developed the SAMR Method in 2006 to understand technology adoption.
SAMR, which stands for Substitution, Augmentation, Modification and Redefinition is complementary to Bloom’s taxonomy. Ruben has observed that the biggest step in integrating technology into your workflow was to make the leap from Augmentation to Modification.
Coincidentally, that is the same struggle I witnessed in my students because Augmentation to Modification coincided with where my students had to transition between logic and evaluation. I see lots of people using AI as an Oracle but relatively few using it as a partner for creation. This is why.
Our labeling systems don’t capture that jump and this matters for how people use technology in the workforce. As Anderson and Krathwohl et al conclude: “In fact, the predominant use of the original [Bloom’s Taxonomy] framework has been in the analysis of curricula and examinations to demonstrate their overemphasis on remembering and their lack of emphasis on the more complex process categories.”
Companies want you to use AI, but they don’t really understand how you should use it. Since most educational assessment stops at the lowest cognitive levels, they aren’t getting any help from their credentialed employees whose labels were granted by reputable institutions of higher learning.
A Magnifying Glass for Context
This ability to do deeper analysis on the fly is why AI will undermine our labeling systems. I no longer need to trust in “appeals to authority.” I can use an AI to analyze the thinking paths of students or potential employees.
I can research a topic and create a much more nuanced picture than even a Harvard professor could do in a 5-minute news segment. I can do this in less than 30 minutes (and could probably figure out how to do it even quicker).
I no longer have to bow to the “authority” of another person because I don’t have time to dig deeper. To borrow from Dr. Martin Luther King, I am at a day when I can evaluate a man for the content of his character, not just the label he carries. AI gives me a magnifying glass to zoom in on context.
This has immense implications for a world dominated by labels for things we don’t think we can understand. It allows me to track connections back to their origins with very little effort. This is now a central lesson of my teaching.
AI allows me to track connections around people. Resumes are lists of labels. They may hint at the person behind them, but no one is that shallow.
To measure the content of someone’s character, I have to dig much deeper than that. Interviews are time consuming and superficial. You’re still trusting in first impressions and just skimming context (if you’re lucky).
AI can help us overcome what AI is destroying. AI can save immense time and allow us to cover more ground and do it quickly than was possible before, even with search engines.
I can get much closer to understanding the content of the character of the person I am hiring. It is possible to get a sense of whether they can do the higher levels of Bloom’s taxonomy. Most post-AI jobs will demand those levels of thinking. As Bowen and Watson point out: “Employers responded to the internet by asking for employees who could do more than just Google an answer, and now they will want to hire graduates who can also do more than just ask AI.”
Those frustrated graduates who have these higher cognitive skills and who can’t get entry level jobs will start up companies that rely on AI to do those expensive entry level jobs from the outset. They will go on to build the products and ideas of the future, but they will do it outside the traditional hierarchical structures of education and business.
The Tangible Work Product
This leaves those of us “selling” credentials in an interesting place. Old-fashioned credentials are going to depreciate. We need much more comprehensive labeling systems.
That’s why I have my students create a Tangible Work Product, for instance. Having a website that demonstrates that they know how to communicate complex information in an understandable way bypasses the opaque labeling of their grade. Their ability to evaluate, analyze, and create is made tangible.
Using AI to get there is a bonus, not a bug. Seeing the reality of their ability to analyze and create tells an employer far more than an “A” in my government class ever could.
Before AI, it was too hard and time-consuming to find and narrow down these information sources. Now, I can just ask Gemini to find out everything it can about a particular candidate and how they think. It doesn’t just tell me that Pat Smith graduated from Houston City College, it tells me what they learned to do there.
Even better, because candidates are building Tangible Work Products in the open, I can ask an LLM to find the Pat Smiths in my area whose specific, demonstrated problem-solving approaches match the exact friction my company is facing.
I am not scraping their resumes. I am discovering their competence. Then, I just reach out to them rather than engaging in the blind man’s bluff of current hiring practices.
This is the future reality of credentialing. Simply having a piece of paper that shows you can get the job will no longer be good enough. Having a demonstration that you have the character and insight to do the job will determine whether I hire you.
If you show how you can work at cognitive levels above the AI’s level, I will let you onto the team. I already know you can operate well beyond today’s entry level jobs because I can see your depth, not just your label. In this scenario, credentialing from schools becomes superfluous.
The Humanities Rebound
That doesn’t mean formal education is dead. Instead, it will need to go back to its roots. Teaching students how to work at the AI-proof levels of Bloom’s taxonomy is hard.
Like all technologies, labels are amoral. Returning to Iacono’s Claude story, the way that they fixed Claude was to give it essentially what amounted to a liberal arts education much broader than technological thrillers. They gave it depth beyond the stereotypes.
This is why AI makes the deeper education implied in the humanities possible. If we can decouple that from the labeling belief that “you’ll never get a job with a Philosophy degree” and start teaching the higher cognitive levels required to understand what philosophy actually means, we’re arming them for success in an AI-world.
Under the old system, the label equals the possibility of employment post-graduation. However, entry-level specialized positions are precisely the kinds of entry-level jobs that AI is automating. Being at the low end of the business hierarchy is also working at the low end of Bloom’s taxonomy. Absent something else, this is a recipe for failure.
Instead of making labels more meaningful, we pivot to network connections. College establishes the lifelong networks that will determine their career paths and prospects, but this biases the system toward elite schools.
Elite schools can charge elite prices because they open the doors to elite networks. This is not a new thing. However, it sustains and deepens the cultural and economic divides we are struggling with today. Elite labels are no more guarantees of quality than any other labels but they do reinforce inequality.
Equity demands that skills should also come into play when determining someone’s career path. AI gives those with skill the ability to demonstrate and share that “credential” in completely new ways, a decisive advantage over those who rely on labels.
The current system fails at this. Labeling someone as inferior is psychologically damaging. Even if they don’t believe it themselves, they will find themselves marginalized because others never get past the label and find the content of their character.
Often these “labels” are more subtle than the color of your skin or your gender even if they serve as proxies for them. Certain people are admitted to the gates of exclusive education and opportunities. Others aren’t.
The difference often comes down to labels. The differences among those groups are far greater than the differences between them. That’s what Dr. King was talking about. AI can show us the way to the world as it is rather than simply the signs that tell us what it is. The only question is whether we will want to use it in this way.