Skip to main content

Table 3 Distribution of type of recommendations among journals providing any own GAI usage guidelines

From: Generative artificial intelligence (GAI) usage guidelines for scholarly publishing: a cross-sectional study of medical journals

A. Journals providing any own guidelines for authors

 

Top SJR ranked journals

(n = 160)#

Random sample of non-top SJR ranked journals (n = 72)#

P value*

Author guidelines, n (%)

 Usage forbidden

0 (0.0)

0 (0.0)

-

 Usage permitted

160 (100.0)

72 (100.0)

 

Type of recommendations among journals permitting GAI use, n (%)

 Language editing

80 (50.0)

44 (61.1)

0.12

 Manuscript writing

97 (60.6)

51 (70.8)

0.13

 Data analysis and interpretation

52 (32.5)

12 (16.7)

< 0.05

 Image generating

89 (55.6)

35 (48.6)

0.32

 Fact-checking

87 (54.4)

39 (54.2)

0.98

 Usage documentation

155 (96.9)

71 (98.6)

0.44

 Authorship eligibility

147 (91.9)

69 (95.8)

0.27

Number of recommendations, median (Q1, Q3)

 Author guidelines (range: 0-7)

5 (3, 6)

5 (3, 6)

< 0.05

B. Journals providing any own guidelines for reviewers

 

Top SJR ranked journals (n = 98)#

Random sample of non-top SJR ranked journals (n = 35)#

P value*

Reviewer guidelines, n (%)

 Usage forbidden

50 (51.0)

16 (45.7)

 0.59

 Usage permitted

48 (49.0)

19 (54.3)

Type of recommendations among journals permitting GAI use, n (%)

 Language editing

19 (39.6)

15 (78.9)

 < 0.01

 Usage documentation

48 (100.0)

11 (57.9)

 < 0.01

Number of recommendations, median (Q1, Q3)

 Reviewer guidelines (range: 0-2)

1 (1, 2)

1 (1, 2)

 < 0.01

  1. Footnote: Raw counts, weighted proportions, and weighted P values were reported. Q1, 25th quantile, Q3, 75th quantile
  2. #Journals providing any own guidelines were included (Own guidelines only and Own & external guidelines in Table 2)
  3. *Differences between the two groups of journals were analyzed using Wilcoxon Rank Sum tests, t-tests, or Chi-Square tests
  4. The percentage was calculated using the number of “usage permitted” journals as the denominator