Home Sports Talk

AI Grading Future - Consistency , Corruption and Payoff ???

FredJRIFredJRI Posts: 390 ✭✭✭
edited May 11, 2024 11:56AM in Sports Talk

Sometimes I like to think out of the box ,... way out!!! lololol

I was curious as to how many people (eyes and hands) actually saw a card(s) submitted for grading prior to reaching the grader so I contacted PSA and asked them.
They couldn't answer specifically how many, but did state apparently several departments. or people would physically see the submission prior to reaching the actual grader. That is good in a way:)
Let's suppose it went directly to the grader and the person submitting the card(s) slipped a few 'C'-notes into the box. It could be possible that the grader(s) grab the 'C'-notes' and in return graded the card one or two grades above what they normally would. Hence an 8 or 9 in lieu of a 6 or 7.
The 'possibility' of corruption and dishonesty exists whenever the human factor is involved.
After all card grading is mostly subjective, I presume, to a certain point within specified guidelines.

I'd be willing to bet a 91 Wild Card, Brett Farve RC "ungraded" that somewhere someplace along the line a grader was essentially paid off to increase the actual grade of a card. Maybe not specifically at PSA or BGS ..but at some grading company.
Maybe by a friend, neighbor someone who knew that person and told them to keep a look out for a particular submission.

That leads me to my next thought: "AI GRADING" ... consistent, full proof, human less, Artificial Intelligence (AI) sport card grading. Parameters and variances are programed, set and locked. The grade is what AI says it is. Absolutely, no human factor or judgement involved. Consistent AI card grading at its best. hehehehehe

In this age of technology, we live in AI will change the way people work, learn, travel, get health care, and communicate with each other's. AI development is as fundamental as the creation of the microprocessor, the PC, the Internet and the mobile phone.
AI technology is being used to fly F-16 fighter planes with superb results, already outperforming human flown F-16.

So my question is not how .. but when do you think "AI card grading" come into full fruition ?

Will your 9's and 10's need regrading ????

FredRI ( Out of the Box )

Comments

  • FredJRIFredJRI Posts: 390 ✭✭✭

    Will "AI Sport Card Grading" be the next thing of the future ?

  • MsMorrisineMsMorrisine Posts: 33,069 ✭✭✭✭✭

    this corruption talk may be a bannable offense

    Current maintainer of Stone's Master List of Favorite Websites // My BST transactions
  • bgrbgr Posts: 1,687 ✭✭✭✭✭

    With respect to card grading? Application of technology is happening. Whether it's truly AI is a stretch. "AI" is becoming the collective term for anything related to software algorithms.

    First, you have to, at least, break it down into the two main forms... statistical analysis (ie. detecting anomalies / clustering), and machine learning (meaning a neural network -- used to be called "modeling").

    What are those two technologies used for?

    • Modeling is used to map an object to a pattern. Is this a ball? Is it a wizard? Is it a GEM MT 10 card. Here you just need a representative image as the "perfect example" of the card, and then you can identify the correlation coefficients for any particular metric you want to evaluate.

    this would be a useful technique for determining centering, coloring, corner and edge quality. measuring the subject against the attributes of a perfect example.

    this would not be useful for identifying if a card is a 7 vs. 8 so much as it would tell you this is a 10 or this is not a 10. there is an interesting study on the human face and beauty which is used to explain the franken-image. ie. taking the most beautiful nose from one person, eyes from another, cheek bones from yet another... so on... then composite them into an image. you do get some good ones, but you also get a lot of trolls. you could think of this as eye-appeal in cards. There is a lot of work going into non-interactive qualitative analysis of compositional elements, but this is frontier work...

    • Clustering is used to assemble things into buckets based on how similar, or dissimilar they are. Think variance.

    this would be useful for identifying surface issues and classifying attributes.

    Clustering, which usually refers to k-means statistical clustering would be the primary technique for evaluating the existing examples against a dataset. ie. this corner looks like this category of corners, and this category of corners is from cards graded as a 6, so these corners are likely a 6. But just because the card is a 6 doesn't mean the corners were a 6. The corners could have been a 9.

    Humans are great at over-simplifying things. You know. "I had a great conversation with a chat bot, so AI is super powerful. It can easily grade cards.". or. "AI made a portrait of Santa Claus riding a hotdog. It gets me! No reason it can't tell me if this is a 5 or a 9.".

    There are a few companies that say they are doing "AI grading", but what that means... it's not always clear. TAG is the big one that went so far as to at least get patents. I am in the minority of people who reviewed their patents I'm sure. It's laughable. The majority of the patents apply to the box the card goes in for images to be captured - technologically ground breaking.... Now we can test Schrodinger's cat. Finally we'll know if it's alive or dead!

    So. I don't think we see automated AI grading any time soon. I added the qualifier because you mention "human less" only to be sure to clarify. I do think that companies are already using image analysis to catch things humans might miss and to provide information to the grader and/or to check the card after the grader has finished. This seems logical and beneficial to all.

  • bgrbgr Posts: 1,687 ✭✭✭✭✭

    @MsMorrisine said:
    this corruption talk may be a bannable offense

    I always put cookies in with my submissions. "If you're not cheating. You're not trying!"

  • ElMagoStrikeZoneElMagoStrikeZone Posts: 497 ✭✭✭

    @FredJRI said:
    Will "AI Sport Card Grading" be the next thing of the future ?

    Not until using lolololololol gets catapulted into the sun.

    Gobble.

  • FredJRIFredJRI Posts: 390 ✭✭✭

    @ElMagoStrikeZone said:

    @FredJRI said:
    Will "AI Sport Card Grading" be the next thing of the future ?

    Not until using lolololololol gets catapulted into the sun.

    lololololololol

  • FredJRIFredJRI Posts: 390 ✭✭✭

    @bgr said:

    @MsMorrisine said:
    this corruption talk may be a bannable offense

    I always put cookies in with my submissions. "If you're not cheating. You're not trying!"

    @bgr said:

    @MsMorrisine said:
    this corruption talk may be a bannable offense

    I always put cookies in with my submissions. "**********If you're not cheating. You're not trying!"


    I believe Tom Brady may have said that along with many others.

    lololololol

  • FredJRIFredJRI Posts: 390 ✭✭✭

    How can one get banned from PSA.. curious? ..have rules and guidelines?
    Is there an area where you can find a list all members, dated joined, # posts, etc?.
    Who are the most knowledgeable members .. who are the PITA members?

  • burghmanburghman Posts: 952 ✭✭✭✭

    Forum rules, posted at the top of the trading cards forum: https://forums.collectors.com/discussion/898536/forum-rules

    Jim

  • estangestang Posts: 1,322 ✭✭✭

    The closest to AI grading is this company:

    https://taggrading.com/

    Enjoy your collection!
    Erik
  • FredJRIFredJRI Posts: 390 ✭✭✭

    @burghman said:
    Forum rules, posted at the top of the trading cards forum: https://forums.collectors.com/discussion/898536/forum-rules

    Thank you burghman ..svaed me some search time.

  • FredJRIFredJRI Posts: 390 ✭✭✭

    @estang said:
    The closest to AI grading is this company:

    https://taggrading.com/

    Thank you sir, it's begun.
    It's the sign of the time, I predict that by 2035, if man is still alive, lololol

  • FredJRIFredJRI Posts: 390 ✭✭✭

    @MsMorrisine said:
    this corruption talk may be a bannable offense

    WHY.. explain ???

  • burghmanburghman Posts: 952 ✭✭✭✭

    Those of us who’ve been here for awhile have seen people banned for various reasons. I think suggesting corruption at PSA is getting pretty close to the line. Rule 4 kinda touches on it:

    Rule 4) Anyone making libelous or defamatory remarks concerning any individual, company, or other entity will no longer be allowed to post.

    Jim

  • LandrysFedoraLandrysFedora Posts: 2,128 ✭✭✭✭✭

    This is not a subject for a psa (or any other grading company run forum), @burghman just posted this perfectly above.

  • FredJRIFredJRI Posts: 390 ✭✭✭

    The subject is really AI Grading Future ... changed it , Happy now ?

    The prelude or premise is us just one reason why it may occur.( AI Grading )
    I never specifically said it occurred at PSA, BGS or any specific company.
    However, I'd be willing to leave here on my own if anyone " could prove that it didn't occur somewhere !!!"

    No reason for anyone to get their panties tied up in a knot.
    I assume most people here are grown adults and can discuss controversial topic.. no?

    If I'm wrong than 'flag' me and let the powers to be decide.
    I'm a new member and didn't come her to create chaos, dissention or trouble.
    I thought it was a legit question which people could discuss and express opinions on ...sheeesh !!

    FredJRI

    PS: Sorry if I offended anyone or broke rules, that was not my intent !!!!!

  • FredJRIFredJRI Posts: 390 ✭✭✭

    @estang said:
    The closest to AI grading is this company:

    https://taggrading.com/

    They will be bought out soon ..my prediction !! lolololol

  • jay0791jay0791 Posts: 3,527 ✭✭✭✭

    I always put payday candy bars in the box. Hoping Hint Hint for all 10's and thus a big payday.

    Collecting PSA... FB,BK,HK,and BB HOF RC sets
    1948-76 Topps FB Sets
    FB & BB HOF Player sets
    1948-1993 NY Yankee Team Sets
  • FredJRIFredJRI Posts: 390 ✭✭✭

    Oh no,100 Grand works btter ,...sometimes! lololol

Sign In or Register to comment.