The Adoption of Artificial Intelligence in Firms

Annex H. Correlations between AI intensity and the perceived helpfulness of selected public sector services

Table H.1. Correlations between AI intensity and how helpful 840 enterprises across G7 countries find selected public sector services, 2022-23

Scores take values from 1 (very helpful) to 4 (not helpful at all)

	Information or advice on business use case	Information or advice on the expected rate of return on AI investment	Information on available and reliable AI vendors	Information on available and reliable sources of private-sector advice	Certification or accreditation schemes for AI providers	Information on current or forthcoming regulation
Number of AI uses	-0.098***	-0.036**	-0.026*	-0.043***	-0.055***	-0.056***
	(-0.01)	(-0.01)	(-0.01)	(-0.01)	(-0.01)	(-0.01)
Number of AI obstacles	-0.054**	-0.054**	-0.035	-0.056**	-0.062**	-0.039*
	(-0.02)	(-0.02)	(-0.02)	(-0.02)	(-0.02)	(-0.02)
Size	0.098	-0.02	0.034	0.102	0.121*	0.008
	(0.05)	(0.06)	(0.06)	(0.05)	(0.06)	(0.06)
Age	0.02	-0.002	-0.011	-0.009	0.016	-0.006
	(-0.01)	(-0.01)	(-0.01)	(-0.01)	(0.01)	(-0.01)
R2	0.141	0.035	0.041	0.069	0.081	0.054
Number of observations	836	836	836	836	836	836
Memo: Average score	1.79	1.85	1.82	1.89	1.97	1.79

Note: The table presents the coefficients of OLS regressions for each service. Robust standard errors are in parentheses, where ***, ** and * represent statistical significance at 1%, 5% and 10% levels, respectively. All regressions are run with country fixed effects. The industry and country dummies are mostly statistically insignificant, suggesting that there is little to no systematic variation across countries and industries. The memos at the foot of the table show the average score on each item in the total sample of active AI users.

Table H.2. Correlations between AI intensity and the score on how helpful 840 enterprises across G7 countries find selected public initiatives, 2022-23

Scores take values from 1 (very helpful) to 4 (not helpful at all)

	Investing in education in fields related to AI	Investing in retraining for employees who work with AI	Improving understanding of AI among government officials	Gathering and publishing administrative public databases	Promoting a competitive AI vendor market
Number of AI uses	-0.078***	-0.014	-0.056***	-0.079***	-0.077***
	(-0.01)	(-0.01)	(-0.01)	(-0.01)	(-0.01)
Number of AI obstacles	0.007	-0.026	-0.061**	-0.059**	-0.045*
	(0.02)	(-0.02)	(-0.02)	(-0.02)	(-0.02)
Size	0.013	0.154**	0.054	0.03	0.032
	(0.06)	(0.05)	(0.06)	(0.06)	(0.06)
Age	-0.016	0.022	-0.009	-0.021	-0.009
	(-0.01)	(0.01)	(-0.01)	(-0.01)	(-0.01)
R2	0.101	0.03	0.063	0.091	0.073
Number of observations	836	836	836	836	836
Memo: Average score	1.83	1.71	1.97	1.99	1.93

Note: The table presents the coefficients of OLS regressions for each service. Robust standard errors are in parentheses, where ***, ** and * represent statistical significance at 1%, 5% and 10% levels, respectively. All regressions are run with country fixed effects. The industry and country dummies are mostly statistically insignificant, suggesting that there is little to no systematic variation across countries and industries. The memos at the foot of the table show the average score on each item in the total sample of active AI users.