Behavioral Tells by “Authors” Using GenAI for Chapter Fakes

The latest scam in academic publishing involves the uses of generative AI to create “fake” articles and chapters for publication. There are ways to head off such fakes, through manual and computational means of assessing submitted informational contents. There are also behavioral tells as well—specifically, how the “authors” behave during the publication process. This process involves the initial query (article or chapter proposal), the submittal, the reviews, and publication. This slideshow is a follow-on slideshow to the initial one linked above.

1. Behavioral Tells by
“Authors” Using GenAI
for Chapter Fakes

2. Overview
•The latest scam in academic publishing involves the uses of
generative AI to create “fake” articles and chapters for publication.
There are ways to head off such fakes, through manual and
computational meansof assessing submitted informational
contents. There are also behavioral tells as well—specifically,
how the “authors” behave during the publication process. This
process involves the initial query (article or chapter proposal), the
submittal, the reviews, and publication. This slideshow is a
follow-on slideshow to the initial one linked above.
2

3. Professional
Chops and
KSAs
3

4. Professional Chops and KSAs
•One behavioral tell of academic deception involves overclaiming
professional chops and knowledge, skills and abilities / attitudes
(KSAs) in their field.
•Many will mention an academic affiliation but use a generic email
account (gmail, yahoo). [Emails linked to the school may show
their status, such as @student.educationalinst.edu.]
•They may claim degrees that they do not have.
4

5. Professional
Depth
5

6. Professional Depth
•There can be an unbridgeable gap between an amateur mental
model vs. an expert conceptual model.
•An amateur may go for simplicity. A professional may go for
complexity.
•An amateur may go for the corner-cutting and the easy-way-in. A
professional would rather ask difficult questions, pursue tough
ambitions and objectives, and apply complexity to their work.
•An amateur may go for the easy understanding. An expert may
better apply critical thinking and healthy skepticism. They can
wield qualifiers more accurately and in a more nuanced way.
6

7. Superficiality vs. Cutting-Edge Innovations
•A generative AI can provide a general overview, but it cannot fake
expertise in any convincing way. The ideas are not particularly
innovative or cutting-edge. There is a lot of superficiality. Many of
the chapter proposals are ragged outlines of ideas.
•Those using genAI to create false contents will show unusual
gaps in their knowledge. They will not have the ability to engage
complexity. They cannot adapt in a way that seasoned individuals
might.
7

8. Often Written Summaries Only
•The works may be reviews of the literature only.
•There is no primary research, only secondary and tertiary
research.
•There is no original data.
•There is no problem solving.
•There is no innovation.
•There is no new technique.
•There is no fresh application of techniques with technologies.
8

9. Often Written Summaries Only (cont.)
•There is a non-obvious value proposition per se.
•In some cases, perhaps, there is the hint of a “homework
assignment” in that a student may have read the work (or its
abstract) or read a genAI summary of a work.
9

10. Swiped Data, Swiped Visuals, Swiped Credit
•Such “authors” will swipe others’ data, others’ visuals, and others’
credit.
10

11. Access to
Cited Sources
for the Review
of the
Literature
11

12. Access to Cited Sources
•Some early warning signs of possible genAI content coming down
the pike is a chapter proposal with DOIs that do not exist.
•There may be cited sources that involve esoteric and subscription
databases that the authors allegedly have access to but likely not
given the focus of their institution of higher education. (It does
take savvy and resources and money to grow expertise of various
kinds.)
12

13. Current
Knowledge
13

14. Current Knowledge
•Maintaining a current sense of knowledge in a discipline or
industry takes a lot of effort and resources. It takes a lot of reading
and analysis.
•Those who use generative AI to create published works…lack this
level of investment. They lack the initiative.
•They conflate the writing of text prompts to genAI tools to actually
engaging in research design, thorough reviews of the literature,
analysis, conducting research, writing it up, and going through
multiple and tough review processes.
14

15. Current Knowledge (cont.)
•AI cannot take the place of an individual’s drive, risk-taking, formal
learning, developed skillsets, ambitions, and energy. The tool may
short-circuit the individual’s actual investment in their actual
professional development.
•AI cannot replace professional relationships and ties. It cannot
bridge interdisciplinary collaborations.
•It cannot replace a sense of professional ethics and professional
pride.
•It cannot create resilience where an individual is not willing to
train into that sense of endurance through hardship.
15

16. Current Knowledge (cont.)
•It cannot create an angle that a researcher or a research team can
exploit unless that individual or team has the skills to act on an
insight or idea or path.
•Those who go with genAI for their research and writing and data
needs are ceding ground already. They have chosen to not pursue
work that is relevant to the field.
16

17. Professional
Affiliations
and Standing
17

18. Professional Affiliations
•A professional affiliation refers to a formal tie to a particular
institution of higher education or other workplace.
•It often means that there are some credentials to the individual’s
name.
•It often means that the individual has gone through some type of
professional vetting.
•If often means that the individual has some oversight to their
work.
•It often means that the individual has professional colleagues with
whom to collaborate and to whom they may want to impress.
18

19. Professional Standing
•Those with professional standing may have affiliations with
professional organizations.
•They may have professional ethical standards that they attest and
adhere to.
•They have worked to earn (or test into) particular credentials.
19

20. No Affiliations? No Professional Standing?
•It is a red flag when the aspiring author / researcher has no
positionality in relation to the topic. They have never studied in the
field. They do not have a guiding professor. They are self-
appointed “experts.” (Dunning-Kruger effect, anyone?)
•Some aspiring researchers / authors may propose various works,
but they do not have a track record of either professional
affiliations or professional standing. They do not have a track
record of successful work.
20

21. Timings and
Prolific
Generic
Production
21

22. Timings
•Like many fraudsters who find that they’re on to a “good thing” (in
terms of deceptions and rewards), those who use genAI to
propose chapter drafts and to write actual chapters…propose a
large number of works. They think that they have found their way
to a reputation-generating machine.
•They have something to play for.
22

23. Prolific Generic Production
•The gen AI-supported would-be published authors often show fast
productivity. They have one idea after another. Their proposals all
look of-a-piece, with a general outline and source citations (some
hallucinated).
•Or the work smacks of student readings, augmented with genAI
contents.
•The publication proposals do not show personal obsession or
interest. They do not show background. There is a sense that the
author is piggy-backing on others’ ideas.
23

24. Prolific Generic Production (cont.)
•The would-be published authors communicate a sense of
costlessness to their work. (After all, they just put in prompts and
clean up what the genAI produces…and they are suddenly a
published academician. Sometimes, genAI can change facts on
the ground in terms of their “achievements.”)
24

25. Professional
Track
Records
25

26. Professional Track Records
•To acquire a particular skillset requires many years of experience.
•A person just out of high school will not have a complex
knowledge base and set of skills to achieve particular outcomes.
26

27. Staking Out a Professional Position
•Those in academia stake out a position or two or three on which to
build their KSAs.
•They opt-in to particular disciplines because of some experience
or ideas.
•They learn and develop over time. They develop the KSAs to
survive in that space.
•They take pride in the professional practices and ethics of the
field(s).
•They invest real time, real energy, real learning…into that work.
They show up.
27

28. Staking Out a Professional Position (cont.)
•Those who go to deceptions will often try to build out a whole
career from such. They try to build out career opportunities from
“their” publications.
28

29. Attitudes and
Entitlements
29

30. When Thwarted…
•When challenged, some go silent.
•Some push back. They feel entitled to make their false claims.
•Some going to harassment emails. They think they have a right to
engage in fraud, and they are annoyed that they’ve been called
out. (It’s simple enough to block these.)
•Some want a chance to revise…to hide the hallucinated
information and sourcing. They argue that the ideas have merit
even if genAI did the work.
30

31. No Actual Investment in the Work
•Many do not think that they will be caught out. They do not
understand what tools the system has to identify fraudsters.
•They are not concerned that they will build their curriculum vitae
with false entries.
•They do not understand that publishers will retract works.
Retractions can irreparably harm a career in academia. They can
be career-ending (and often are in top-tier institutions of higher
education).
•Lower-tier IHEs have individuals who may create an authorizing
environment that enables deceptions.
31

32. No Actual Investment in the Work (cont.)
•To survive in a field requires a durable competitiveness, not a
facades, not one-offs. It requires actual KSAs.
32

33. The Limits
of
Behavioral
Tells
33

34. The Limits of Behavioral Tells
•That people can be shameless and uninhibited is not news.
•In academic publishing, there are occasional “smoking guns”
which are a clear tell. (Hallucinated sourcing is one.) In other
cases, there are not. The issues are more complex, more subtle.
Perhaps the data is faked. Perhaps the writing piggybacks on
others’ work in unacknowledged ways.
•Behavioral tells are assessed along with a set of other
observations…to come to a decision about a work.
•No one just goes, “Got me!”
34

35. Communications
and Contact
35

36. Communications and Contact
•Dr. Shalin Hai-Jew
•[email protected]
36