Data Processing Back in the Day

Found these computer tapes in my office from back in the day when data sets were fed into mainframe computers on them.

The Godot Recession: We're Still Waiting

It's a bit boring - I know - but every week we examine the latest entrails of data on California new weekly claims for unemployment insurance for signs of a recession. And every week, we don't see those signs. In the labor market, things are pretty much as they were in the pre-pandemic boom period. The latest data for the week through Jan. 28 continue to tell that story.

As always, the latest claims data are at https://www.dol.gov/ui/data.pdf.

A recent Ohman cartoon says it all:

AI Checkers (Classifiers) Aren't There Yet - Part 2

On Friday, we posted about the use of artificial intelligence programs and their higher ed implications.* We noted that unlike conventional plagiarism - where programs to determine where an essay came from such as Turnitin are pretty reliable - programs to tell you whether something was written by a machine are not all that good.

Nonetheless, there are rules being made or on the books which would seem to ban handing in machine-made term papers. But some rewriting may be needed The UCLA rule - written before AI became an issue states:

Plagiarism includes, but is not limited to, the use of another person’s work (including words, ideas, designs, or data) without giving appropriate attribution or citation. This includes, but is not limited to, representing, with or without the intent to deceive, part or all of an entire work obtained by purchase or otherwise, as the Student’s original work; the omission of or failure to acknowledge the true source of the work; or representing an altered but identifiable work of another person or the Student’s own previous work as if it were the Student’s original or new work.

Unless otherwise specified by the faculty member, all submissions, whether in draft or final form, to meet course requirements (including a paper, project, exam, computer program, oral presentation, or other work) must either be the Student’s own work, or must clearly acknowledge the source.

Source: https://deanofstudents.ucla.edu/individual-student-code#c_prohibited_behavior (Section 102.01c)

I have underlined the word "person's" in the item above as something that needs to be modified to make it clear that the copying ban isn't limited to copying from a person.

Meanwhile, the UCLA Anderson School came up with what appears to be an interim statement on this issue:

I just wanted to send along encouragement for faculty and students to talk openly about how, when, and why we may use ChatGPT. As with any advance in technology, we want to be proactive in making sure we use them ethically and as a tool to augment the learning that takes place at Anderson.

 

For students, I want to remind you that the use of ChatGPT still remains an outside source and not original work and as such, should be cited whenever it is used for any assignment and should not be used when original work is requested by your instructor. This is an exciting time and don’t be hesitant to discuss the use of ChatGPT with your instructor before using it for any assignment.

 

For faculty, there are lots of exciting challenges here for us to ask interesting and original assignment questions to push students to go beyond what such a tool can currently produce. I’d encourage discussion and reflection with other instructors and your students on this as we continually seek to improve education...

PS: Now you’ll have to ask yourself whether this email was written by ChatGPT! :-)

Source: Email distributed to faculty last Wednesday by the senior associate dean for MBA programs.

====

Yours truly should note that what such guidelines, whether from particular schools or issued campus-wide, should make clear is that instructors are looking for original work. Technically, cutting and pasting a collage of other people's work (or a machine's work) is not plagiarism if footnotes or other references are provided. But it does not show the ability of the student to compose something.

====

*http://uclafacultyassociation.blogspot.com/2023/02/ai-checkers-classifiers-arent-there-yet.html. 

A sign to remember

Yours truly took a photo of this less-than-helpful campus sign c2004. But he remembers it.

Student-Worker Strike Repercussions - Part 6

In past postings, we have looked at the repercussions from the student-worker strike in terms of the university reaction, i.e., budgets, hiring, etc. As might be expected, there were repercussions on the union side that are still being felt and could eventually express themselves through the union political process. See below from the Daily Bruin:

Catherine Hamilton  2-2-23

Rank-and-file members of several United Auto Workers unions expressed dissatisfaction with issues of marginalization, bullying and suppression they said they experienced during internal negotiations to ratify a new contract with the University of California. On Dec. 23, UAW Local 2865 and Student Researchers United-UAW separately ended their nearly six-week long strikes and ratified new contracts with the UC, with nearly 62% and 69% of rank-and-file members voting yes on the agreements, respectively. Agreements for all three unions, as UAW Local 5810 reached an agreement Dec. 9, included pay increases, increased childcare benefits and anti-harassment clauses. But multiple units at UC Santa Cruz, UC Merced and UC Santa Barbara rejected their proposals by large margins, voicing discontent with their bargaining team members’ decisions to concede on lower wages and cut representative clauses.

Despite having opportunities to speak throughout the negotiation process, many rank-and-file members felt their representatives were not listening to their concerns, said Adam Moore, a pathobiology doctoral candidate at UC Davis and member of UAW Local 2865 and SRU-UAW. Internal bargaining sessions to decide on proposals were open to all members of the unions, and a caucus was held after for members to express desires, support and concerns to the bargaining units, he said. “It seemed like it didn’t matter how many times or how many different people shared their really personal and sometimes traumatic stories,” Moore said. “My impression was that their minds had already been made up, and caucus was just a space for us to talk to ourselves.”

Janna Haider, a history doctoral student at UCSB and member of the UAW Local 2865 bargaining team, said she believed the contract would not make enough changes to ease the rent burden at her campus. She added that she was against many decisions ultimately made by the bargaining team and organized people to vote no against the contract. Even before the strike began, Haider said union meetings felt more like presentations from the bargaining team on decisions already made rather than a collaborative space to discuss. This was especially felt in the setup of the strike itself, she added.

“The statewide UAW infrastructure insisted on 20 hours a week (of picket duties) … (and) really wanted to prioritize the in-person picket line, as if standing around outside is the thing that indicates how strong the strike is,” she said. “Santa Barbara was the only campus that had an infrastructure for an online picket for workers who are disabled (or) who are parents who live far away from campus.”

Nick Geiser, a physics doctoral student and member of the SRU-UAW bargaining team, said in an emailed statement that the union allowed opportunities for democratic input through surveys, votes and open bargaining sessions. At UCLA, bargaining team members spoke directly with their peers, held collaborative discussions and engaged rank-and-file members throughout the process, he said, adding that the units recognize there is still work to be done.

“We of course agree that the fight against rent burden is not over, and as a union we are taking our political fight for housing justice to the state legislature in Sacramento while gearing up for our next round of negotiations in 2025,” he said in the emailed statement. However, Haider said there was heavy pressure from many members of the bargaining team to finish negotiations quickly, resulting in a final contract that did not immediately raise wages and instead forced all workers to wait until the spring to receive new levels of pay...

Full story at https://dailybruin.com/2023/02/02/a-month-after-new-contract-uaw-members-concerns-about-bargaining-process-persist.

Just Taking a Walk

As long as we're worrying about artificial intelligence writing student essays - see our previous post - why not worry about robots taking over the campus entirely?

Or direct to https://www.youtube.com/shorts/DyCKXdrnEEA.

The ARTEMIS robot out for a stroll from: 

https://twitter.com/UCLACommute/status/1621276580365627392.

AI Checkers (Classifiers) Aren't There Yet

DALL-E: Create a picture of a female professor trying to determine
whether a student or an artificial intelligence program
wrote an essay in renaissance painting style.
(Four choices were generated by DALL-E.)

There is lots of fretting in academia about students using AI to write essays and term papers. Below are some developments.

From The Verge: OpenAI, the company behind DALL-E and ChatGPT, has released a free tool that it says is meant to “distinguish between text written by a human and text written by AIs.” It warns the classifier is “not fully reliable” in a press release and “should not be used as a primary decision-making tool.” According to OpenAI, it can be useful in trying to determine whether someone is trying to pass off generated text as something that was written by a person.

The tool, known as a classifier, is relatively simple, though you will have to have a free OpenAI account to use it. You just paste text into a box, click a button, and it’ll tell you whether it thinks the text is very unlikely, unlikely, unclear if it is, possibly, or likely AI-generated...

OpenAI also says in its tests the tool labeled AI-written text as “likely AI-written” 26 percent of the time and gave false AI detections 9 percent of the time...

Full story at https://www.theverge.com/2023/1/31/23579942/chatgpt-ai-text-detection-openai-classifier.

====

From the above-mentioned press release:

...Limitations

Our classifier has a number of important limitations. It should not be used as a primary decision-making tool, but instead as a complement to other methods of determining the source of a piece of text.

The classifier is very unreliable on short texts (below 1,000 characters). Even longer texts are sometimes incorrectly labeled by the classifier.

Sometimes human-written text will be incorrectly but confidently labeled as AI-written by our classifier.

We recommend using the classifier only for English text. It performs significantly worse in other languages and it is unreliable on code.

Text that is very predictable cannot be reliably identified. For example, it is impossible to predict whether a list of the first 1,000 prime numbers was written by AI or humans, because the correct answer is always the same.

AI-written text can be edited to evade the classifier. Classifiers like ours can be updated and retrained based on successful attacks, but it is unclear whether detection has an advantage in the long-term.

Classifiers based on neural networks are known to be poorly calibrated outside of their training data. For inputs that are very different from text in our training set, the classifier is sometimes extremely confident in a wrong prediction...

Full news release: https://openai.com/blog/new-ai-classifier-for-indicating-ai-written-text/.

===

In short, you can't rely on the classifier to make an accusation of plagiarism, if that is what you call using AI to write an essay. I suspect at this point that if an instructor assigned a term paper and insisted on having footnotes and references, currently available AI programs would have a problem. But eventually even that might not be adequate protection. Since the "I" in AI is something of a misnomer, it may include false information since it is relying on text on the Internet, and there is a lot of false information on the Internet. An essay with perfect grammar and containing false information would be a pretty good indicator of reliance on AI at this stage. The classifier is as unlikely to recognize false information as is the original AI program that wrote the essay. But presumably the instructor who assigned the essay can make that determination.