Readiness

Assessment
Methodology
A Tool of the Recommendation
on the Ethics of Artificial Intelligence
Published in 2023 by the United Nations Educational, Scientific and Cultural Organization
7, place de Fontenoy, 75352 Paris 07 SP, France
© UNESCO 2023

This publication is available in Open Access under the Attribution-ShareAlike 3.0 IGO (CC-BY-SA 3.0 IGO) license
([Link] By using the content of this publication, the users accept to be bound
by the terms of use of the UNESCO Open Access Repository ([Link]
The designations employed and the presentation of material throughout this publication do not imply the expression of
any opinion whatsoever on the part of UNESCO concerning the legal status of any country, territory, city or area or of its
authorities, or concerning the delimitation of its frontiers or boundaries.
The ideas and opinions expressed in this publication are those of the authors; they are not necessarily those of UNESCO and
do not commit the Organization.
Cover photo: SeventyFour/[Link]
Designed and printed by UNESCO
Printed in France
[Link]
Readiness
Assessment
Methodology
A Tool of the Recommendation
on the Ethics of Artificial Intelligence
Table of Contents

Foreword 5

I. Background 6

II. Guidance for filling out the RAM 8

III. Questions 9

1. General Questions 9
1.1 Does your government currently have plans to implement the UNESCO
Recommendation on the Ethics of AI, through establishing a national AI Ethics
Commission or through other measures? 9
1.2 Has your government done any prioritization regarding the specific sectors
that would benefit from government intervention (such as regulation, strategies
or guidelines, etc) regarding AI? 9
1.3 Does your government inform the public when they are subjected to the use of AI
systems that profile or make decisions about them in the provision of public services? 9
1.4 Which ministry/ies are responsible for AI governance?
Please elaborate on the composition of the teams. 9
1.5 Overall, what are the most important challenges to developing
AI regulations and policies in the country? 9

2. The Legal Dimension 9

2.1 Background 9
2.2 Indicators for the legal dimension 10
2.2.1 AI policy and regulation 10
2.2.2 Data protection and privacy laws 11
2.2.3 Data sharing and accessibility 11
2.2.4 Procurement laws and policies 12
2.2.5 Freedom of information acts/access to knowledge acts 12
2.2.6 Due process and accountability 13
2.2.7 Online safety and integrity of speech 13
2.2.8 Public sector capacity 13

3. The Social/Cultural Dimension 14

3.1 Background 14
3.2 Indicators for the Social/Cultural dimension 14
3.2.1 Diversity, inclusion and equality 14
3.2.2 Public engagement and trust 15
3.2.3 Environmental and sustainability policies 16
3.2.4 Health and social well-being 16
3.2.5 Culture 17

3
 4. The Scientific/Educational dimension 18
4.1 Background 18
4.2 Indicators for the scientific/educational dimension 18
4.2.1 Research and innovation 18
[Link] R&D expenditure 18
[Link] Research output 18
[Link] Ethical AI research 19
[Link] AI talent 19
[Link] Innovation output 19
4.2.2 Education 19
[Link] Education strategy 19
[Link] Education infrastructure 20
[Link] Curriculum content 20
[Link] Educational attainment 20
[Link] Public access to AI education 20

5. The Economic Dimension 21

5.1 Background 21
5.2 Indicators for the economic dimension 21
5.2.1 Labour markets 21
5.2.2 Intermediate consumption 21
5.2.3 Investments and output 21

6. The Technical and Infrastructural Dimension 22

6.1 Background 22
6.2 Indicators for the technical dimension 22
6.2.1 Infrastructure and connectivity 22
6.2.2 Applied standards 23
6.2.3 Computing capabilities 23
6.2.4 Statistical performance 24

Annex 26
Foreword

We have officially entered the Age of Artificial Intelligence. The world is now set to change at a pace not seen in decades, even
centuries. AI-based tools and applications make our lives easier, smoother, and richer. They help us move efficiently, get informed,
get credit, get a job, and get our taxes done.

But in its current form, AI reproduces and amplifies many of the social challenges we face.

We need to decide what lies beyond. This is not a technological discussion. It is a societal one, about the world we want to live
in. To shape the technological development of AI, we need effective governance frameworks underpinned by the ethical and
moral values we all hold dear.

That is why UNESCO developed the Recommendation on the Ethics of Artificial Intelligence, which 193 countries have adopted to
make sure AI delivers fair, sustainable, and inclusive outcomes. The Recommendation sets out the values and principles based
on protecting human rights and dignity and translates them into specific policy areas and recommendations for governments,
premised on the belief that self-regulation, which has until now remained the norm, is insufficient. We need capable governments
that protect the rule of law online. We need public and private developers that are accountable for putting people — not profits
or geopolitical considerations — first.

UNESCO’s policy approach recognizes that countries are at different stages of AI development, so we are delivering targeted
policy support for different countries. Despite strong regulatory action in some regions, the institutional shape and capacities of
governments diverge significantly.

This is why Member States asked UNESCO to develop a Readiness Assessment Methodology to assist them in “identifying their
status at specific moments of their readiness trajectory along a continuum of dimensions”. The Readiness Assessment Methodology
(RAM), and the complementary Ethical Impact Assessment tool, were officially launched on 13 December 2022, during the
inaugural Global Forum on the Ethics of AI in Prague, under the Czech Presidency of the European Union. We are now working
with several countries in Latin America, Africa, Asia, and Europe to implement this methodology.

The RAM encompasses five dimensions: Legal and Regulatory, Social and Cultural, Economic, Scientific and Educational, and
Technological and Infrastructural. Each dimension is broken down into sub-categories containing qualitative and quantitative
indicators and sub-indicators for a cohesive assessment. In addition to providing rich information about the status of individual
countries, the RAM will also provide comparative information for countries to learn from each other.

UNESCO’s Secretariat designed this tool with the strong support of a regionally balanced and transdisciplinary High-Level Expert
Group, ensuring it was informed by a range of diverse perspectives.

Conducting the Readiness Assessment will lead to a country report and a specific roadmap built in partnership with the concerned
countries and UNESCO’s experts to suggest the best paths forward and guide the country, particularly the government, in
building their capacities. This includes strengthening human capital and national institutions and implementing and upgrading
policies and regulatory frameworks to address AI challenges.

When the Recommendation was adopted in November 2021, we collectively dreamed up a vision — one where AI is utilized,
developed, and applied ethically, for the benefit of humanity and our planet. In this, regulation and governance play an essential
role. By leveraging tools such as the RAM, UNESCO hopes to work closely with Member States, supporting them in strengthening
capacities for instituting robust AI regulation. Through the RAM, I believe we can move one step closer to fulfilling this ethical
vision for us all.

Gabriela Ramos
Assistant Director-General for Social
and Human Sciences, UNESCO

5
 I. Background

In November 2021, the 193 Member States of UNESCO signed the Recommendation on the Ethics of Artificial Intelligence,
the first global normative instrument in its domain. The Recommendation serves as a comprehensive and actionable
framework for the ethical development and use of AI, encompassing the full spectrum of human rights. It does so by maintaining
focus on all stages of the AI system lifecycle. Beyond elaborating the values and principles that should guide the ethical design,
development and use of AI, the Recommendation lays out the actions required from Member States to ensure the upholding
of such values and principles, through advocating for effective regulation and providing recommendations in various essential
policy areas, such as gender, the environment, and communication and information.

With these values, principles, and policy areas in mind, the UNESCO Secretariat elaborated a programme for the implementation of
the Recommendation, with the core aim of building national capacities to discharge the actions set out in the Recommendation
and bolster regulatory frameworks.

The Recommendation mandated the development of two key tools, the Readiness Assessment Methodology (RAM) and the
Ethical Impact Assessment (EIA), which form the core pillars of the implementation. These tools both aim to assess and promote
the resilience of existing laws, policies and institutions to AI implementation in the country, as well as the alignment of AI systems
with the values and principles set out in the Recommendation.

The goal of this document is to provide more information on the Readiness Assessment Methodology, lay out its various
dimensions, and detail the work plan for the implementing countries, including the type of entities that need to be involved,
responsibilities of each entity, and the split of work between UNESCO and the implementing country.

Instrument: The Recommendation, in paragraph 49, states: “UNESCO recognizes that Member States will be at different stages
of readiness to implement this Recommendation, in terms of scientific, technological, economic, educational, legal, regulatory,
infrastructural, societal, cultural and other dimensions. It is noted that ‘readiness’ here is a dynamic status. In order to facilitate the
effective implementation of this Recommendation, UNESCO will therefore: (1) develop a readiness assessment methodology to
assist interested Member States in identifying their status at specific moments of their readiness trajectory along a continuum of
dimensions;”.

As such, the RAM is a macro level instrument that will help countries understand where they stand on the scale of preparedness to
implement AI ethically and responsibly for all their citizens, in so doing highlighting what institutional and regulatory changes are
needed. The outputs of the RAM will help UNESCO tailor the capacity building efforts to the needs of specific countries. Capacity
here refers to the ability to assess AI systems in line with the Recommendation, the presence of requisite and appropriate human
capital, and infrastructure, policies, and regulations to address the challenges brought about by AI technologies and ensure that
people and their interests are always at the center of AI development.

The Readiness Assessment Methodology was developed by the UNESCO Secretariat with the guidance of the High-Level Expert
Group (HLEG) on the Implementation of the UNESCO Recommendation on the Ethics of AI. The HLEG is composed of more than
20 renowned experts, acting in their individual capacity. The Secretariat conducted initial desk research in order to unpack the
different dimensions listed in the Recommendation into more concrete indicators.

6
The Readiness Methodology is composed of five dimensions: Legal / Regulatory, Social / Cultural, Economic, Scientific /
Educational, and Technological / Infrastructural. To develop the RAM, the Secretariat produced a zero-draft methodology
along those dimensions, with each dimension containing various indicators. The expert group was divided into five sub-groups,
with each discussing an assigned dimension in detail. The draft was then further refined following feedback from the experts.

Within each dimension of the RAM, questions are grouped under different categories. These categories are then unpacked
further into specific indicators and sub-indicators. These indicators include both qualitative and quantitative indices, setting
the RAM apart from other existing readiness assessment tools. Some quantitative indices are already being measured by and
for other databases, resources or indexes, but the goal of the UNESCO RAM is to combine them with other resources in order to
deduce related conclusions. Countries will always be able to update the information on existing indices as needed, or report if
they are collecting the needed information in another way.

The final output of the assessment will entail a country report that provides a comprehensive overview of the status of the
readiness in the country, summarizing where the country stands on each dimension, detailing ongoing initiatives, and summing
up the state of the art. This report will help identify what institutional changes are required to elaborate or bolster a National
AI Strategy, allowing UNESCO to tailor capacity-building efforts to the specific needs of different countries to ensure the ethical
design, development and use of AI.

One of the goals of the RAM is also to provide Member States with incentives to boost their AI policy in the form of evidence,
prompting them to invest in concrete areas which require further development and encouraging them to collect data when
it is lacking, and in this regard, cross country comparison for specific topics and for the purpose of mutual learning, might be
envisaged.

The findings of the RAM will also be shared on UNESCO’s Global Observatory, which will enable the sharing of best practices and
dialogue between countries. Amid this, it is important to highlight that the aim of this exercise is not to rank countries against
each other, but rather to enable positive exchanges, encourage dialogue on best practices and data to be collected.

The implementation of the RAM will be adapted to the unique circumstances and characteristics of the country, as well as in
accordance to the budget available for the project. In each beneficiary country, the readiness assessment will be conducted by an
independent consultant/ research organization, supported by a National Stakeholder Team comprising a variety of stakeholders,
such as personnel from the UNESCO Secretariat and UNESCO National Commission, as well as representatives from the country’s
government, the academic community, civil society and the private sector, among others.

For more details on the implementation process, please refer to the Annex.

7
 II. Guidance for filling
out the RAM

The goal of the RAM is to identify strengths and gaps of beneficiary countries with regards to the capacity to facilitate the
ethical design, development and use of AI, and how to address these. For example, challenges could come from lack of
resources, capacity, or specific political challenges, each of which may require different institutional responses. These findings will
help UNESCO in developing a useful and unique roadmap for the country. The team should keep this in mind when filling out
the RAM and should aim to provide responses that support these aims.

To evaluate the laws, policies or strategies of the beneficiary country, the RAM typically includes sub-questions under indicators so
as to facilitate a more nuanced assessment of their effectiveness (beyond their existence or non-existence). While these provide a
useful set of criteria, the team should not feel constrained by them and are encouraged to also furnish their responses with other
details that have not been explicitly requested, should they feel that this would provide useful insight into specific strengths, or
challenges that need addressing.

For instance, the RAM requests a particularly large amount of detailed information regarding the National AI Strategy (if it exists)
– for example: whether it includes provisions for monitoring and evaluation or whether budgets have been allocated to address
recommendations. The team is highly encouraged to think about whether these questions are relevant for other questions across
all dimensions of the UNESCO Recommendation, even if they are not explicitly asked.

Equally, while the RAM asks what the most important challenges to developing AI regulations and policies in the country are
broadly, it does not reproduce this question across the specific dimensions. The team should, however, make a conscious effort
to systematically highlight the specific challenges to the development or implementation of effective policies and regulations
where relevant, as this is a particularly important element in helping UNESCO develop the country roadmap.

While the team filling out the RAM should make an effort to provide answers to all questions, UNESCO acknowledges that due to
the comprehensiveness of the tool, it might not always be possible to answer each and every question. Nevertheless, it would be
important to provide sufficient information in order to be able to cover all of the categories under each dimension.

Finally, throughout the assessment, there are many quantitative questions which may be pre-filled by referring to (usually publicly
available) data sets where the relevant data can be found. The corresponding dataset or resource that will be utilized is referenced
for each respective indicator. Should a country not be covered by a particular data set, the team can consider whether alternative,
local data may be used, even if not fully comparable, if this can provide relevant insights.

8
III. Questions

1. GENERAL QUESTIONS
This section contains general questions regarding the country’s AI governance structure and the overall challenges it faces in its
development.

1.1 Does your government currently have plans to implement the UNESCO Recommendation on the Ethics of AI,1 through
establishing a national AI Ethics Commission or through other measures?

1.2 Has your government done any prioritization regarding the specific sectors that would benefit from government intervention
(such as regulation, strategies or guidelines, etc) regarding AI?

1.3 Does your government inform the public when they are subjected to the use of AI systems that profile or make decisions
about them in the provision of public services?

1.4 Which ministry/ies are responsible for AI governance? Please elaborate on the composition of the teams.

1.5 Overall, what are the most important challenges to developing AI regulations and policies in the country?

2. THE LEGAL DIMENSION

2.1 Background
The legal/regulatory dimension (including the capacity to implement and enforce the regulatory frameworks), is a key dimension
addressing the institutional and human capacity of Member States to implement the Recommendation and, more generally,
to face the major societal transformations caused by the increased adoption of AI in all sectors of the economy. The regulatory
framework should include aspects of effective protection, enforcement, redress, and monitoring of potential harms related to
the deployment and use of AI systems. This includes assessing whether Member States have adopted appropriate regulatory
frameworks to ensure the ethical development and deployment of AI, as well as monitoring and evaluation mechanisms for its
implementation and enforcement. This dimension should help monitor the existence and effectiveness of laws and concrete
provisions implementing the Recommendation. In the case of privacy, this may include, for example, ensuring current data
privacy and data protection is not compromised with the deployment of AI systems, assessing regulations to ensure gender
equality, or preventing abuse of dominant market positions.

1 [Link]

9
 2.2 Indicators for the legal dimension

2.2.1 AI policy and regulation

QUALITATIVE [Link] Does your country have a national AI strategy? If not, do you have any legislation or strategy which
has an indirect impact on AI regulation (for example, data privacy or anti-discrimination laws, or a
digital strategy)? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of the national AI strategy been assessed? Please elaborate and provide the link to any
relevant document(s).

[Link].2 Does the national AI strategy or equivalent include an ethical component?

[Link].3 Was the national AI strategy or equivalent created by a diverse team (including men and women,
minorities, etc.)?

[Link].4 Was the national AI strategy or equivalent created with consultation from different stakeholders
(academics, business executives, civil society, etc.)?

[Link].5 Does the national AI strategy or equivalent include references to AI impacts on human rights?

[Link].6 Does the AI strategy or equivalent include a detailed implementation plan?

[Link].7 Is there a dedicated body/working group leading the implementation of the AI strategy or equivalent?

[Link].8 Does the AI strategy or equivalent include budgetary allocations for the measures it recommends?

[Link].9 Does the AI strategy or equivalent require that AI impact assessment be conducted before deployment in
certain domains?

[Link].10 Does the AI strategy or equivalent include provisions for monitoring and evaluation?

[Link] Has your country enacted any binding AI regulation or soft law (for example procurement
guidelines)? If your country has not enacted any AI regulation, is it in the process of enacting such
regulation? Please elaborate and provide the name and link to any relevant document(s), or state
the reasons why such regulation has not yet been enacted.2

Evaluation:
[Link].1 Has the efficacy of this binding AI regulation been assessed? Please elaborate and provide the link to any
relevant document(s).

2 Several policies and regulations may be found on the OECD AI Observatory, a live repository of over 800 AI policy initiatives from 69 countries,
territories and the EU ([Link] The OECD AI Observatory will generally be useful to answer policy-
related questions. The GovTech Dataset from the WorldBank (199 countries) may also be useful: [Link]
dataset/0037889/GovTech-Dataset (see indicators I-17.1 to I-17.7.1).

10
2.2.2 Data protection and privacy laws

QUANTITATIVE [Link] Score of the country on the Cybersecurity Index3

QUALITATIVE [Link] Does your country have a data protection law?4 If not, is your country in the process of enacting such
regulation? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of the data protection law been assessed? Please elaborate and provide the link to any
relevant document(s).

[Link].2 Does the data protection law give users control over their data and allow them to delete it?

[Link].3 Does the data protection law mention a notice and consent framework and specify in which cases it
applies?

[Link].4 Does the data protection law include transparency requirements on data usage?

[Link].5 Does the data protection law include requirements on data minimization?

[Link].6 Does the data protection law highlight cases in which data protection or privacy impact assessment is
required?

[Link].7 Does the data protection law include specific rules for sensitive information (e.g., health data)?

[Link].8 Does the data protection law include enforcement mechanisms and compensation schemes in case of
violation?

[Link].9 Are there different standards of data protection applied for data collected by public vs. private entities?

[Link].10 Is privacy and/or respect for private and family life protected under the data protection law or another
law? Please provide the link to any relevant document(s).

[Link].11 Do you have a data protection entity or data protection officer in your country? If yes, what is their
mandate?

2.2.3 Data sharing and accessibility

QUANTITATIVE [Link] Score of the country on the Open Data Inventory5

QUALITATIVE [Link] Did your country sign the international Open Data Charter?6

[Link] Do you have a national data sharing framework?7 If not, is your country in the process of creating
one? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of the national data sharing framework been assessed? Please elaborate and provide the
link to any relevant document(s).

[Link].2 How does the data sharing framework approach data sharing between the public and private sector,
different geographies, etc.?

3 [Link]
4 [Link]
5 [Link]
6 [Link]
7 See Governance (G): Data Sharing Frameworks: [Link]

11
 QUALITATIVE [Link] Does your country have open government data policies?8 If not, is your country in the process of
adopting one? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of open government data policies been assessed? Please elaborate and provide the link
to any relevant document(s).

[Link].2 Does the open government data policy mention making datasets available and accessible for research?

2.2.4 Procurement laws and policies

QUALITATIVE [Link] Does your country have laws or policies regarding procurement of AI systems or products/services
that include AI components?9 If not, are such laws or policies in the process of being adopted?
Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of these laws or policies been assessed? Please elaborate and provide the link to the
relevant document(s).

[Link].2 Is there a special approval process before purchasing AI systems?

[Link].3 Is there a list of certified vendors with AI-related provisions?

If so:
[Link].4 Does your certification entail only technical dimensions, or both technical and ethical dimensions?

[Link].5 Does your certification include AI systems only, or both AI systems and agents through which AI
technology comes into public systems?

[Link].6 Is certification voluntary or mandatory?

2.2.5 Freedom of information acts/access to knowledge acts

QUALITATIVE [Link] Is there a freedom of information act in your country?10 If not, is such an act in the process of
being adopted, if not, why not? Please elaborate and provide the name and link to any relevant
document(s).

Evaluation:
[Link].1 Has the efficacy of this act been assessed? Please elaborate and provide the link to the relevant
document(s).

[Link].2 Can individuals request information about the way AI systems are used in the public sector under this act?

[Link].3 Are there obligations for parties using and/or sharing data to inform those whose data they are using
and/or sharing?

8 See Capability (C): Open data initiative: [Link]

9 We suggest you first consult the Global Public Procurement Database (GPPD), developed by the World Bank:
[Link]
10 [Link] No need to report the precise rating, only the existence (or non-existence) of a freedom of information act in the
country.

12
2.2.6 Due process and accountability

QUALITATIVE [Link] What is the main law protecting due process rights in your country? If not, is such a law or policy
in the process of being adopted? Please elaborate and provide the name and link to any relevant
document(s).

Evaluation:
[Link].1 Has the efficacy of this law been assessed? Please elaborate and provide the link to the relevant document.

[Link].2 Are there situations in which individuals must be informed that they are interacting with AI systems?

[Link] Is there a law or policy highlighting monitoring, redress, and remedy mechanisms against harms
caused by AI systems? If so, which mechanisms? If not, is such a law or policy in the process of being
adopted? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of this law been assessed? Please elaborate and provide the link to the relevant document.

[Link].2 Can regulators or courts request information about AI systems and their inner workings under this law/
policy?

2.2.7 Online safety and integrity of speech

QUALITATIVE [Link] Is there a framework for notice and take down policies for violating content, such as online hate
speech, misinformation and disinformation? If not, is such a framework in the process of being
adopted? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has there been any assessment of the efficacy of this law or policy? Please provide the link to any relevant
document(s).

[Link].2 In which cases is this framework applied?

[Link].3 Is the responsibility of online intermediaries defined under this law?

[Link].4 What is the responsibility of online inter­mediaries under this law?

[Link] Has your country enacted any law or policy regarding the impact of AI on social media, including
about transparency, misinformation, disinformation and hate speech? If not, is such a framework
in the process of being adopted? Please elaborate and provide the name and link to any relevant
document(s).

Evaluation:
[Link].1 Has there been any assessment of the efficacy of this law or policy? Please provide the link to any relevant
document(s).

2.2.8 Public sector capacity

QUALITATIVE [Link] Is there a government strategy/programme to improve digital skills in the public sector?11

Evaluation:
[Link].1 Is there a Digital Skills programme?

[Link].2 If so, is the Digital Skills programme mandatory for new public employees?

11 Source: World Bank GovTech Maturity Index ([Link]

13
 3. THE SOCIAL/CULTURAL DIMENSION

3.1 Background
This dimension will consider the factors relevant for ethical development and deployment of AI systems including inclusiveness
and social and cultural diversity, public awareness and values relevant to scaling up of ethical AI solutions. The idea is that if the
teams developing and deploying AI systems are very homogeneous it may lead to AI systems not sufficiently reflecting the
complexity and diversity that society is composed of, which means the outcomes generated by AI systems may contribute to
amplification of structural bias. Second, it will address attitudes towards AI technologies, including their public acceptance. It
should also help reveal societal values and preferences in Member States, which lead to certain attitudes towards technologies
and directly impact related societal choices. Some of the issues it aims to address are as follows: First, it will address the issues of
respect for inclusiveness and social and cultural diversity, including the level of gender representation in all stages of the AI life
cycle, as well as the involvement of different communities and minorities that are impacted by the technology. This aspect aims
to address the current gap and underrepresentation of women and minorities in the AI field, including lack of female developers,
researchers, professors etc. This dimension will also ask crucial questions about sustainability and the environment, including on
provisions to address the environmental impacts of AI systems.

3.2 Indicators for the Social/Cultural dimension

3.2.1 Diversity, inclusion and equality

QUANTITATIVE [Link] Gender gap in internet use12

[Link] Rural/urban gap in internet use (individuals)13

[Link] Percentage of male/female tertiary education graduates in STEM programmes14

[Link] Ratio top girls/boys in science or mathematics who expect to work as STEM professionals when they
are 3015

[Link] Science performance difference (boys vs. girls)16

QUALITATIVE [Link] Has your country enacted any law or policy to reduce the digital gender gap17? If not, is such a law
or policy in the process of being adopted? Please elaborate and provide the name and link to any
relevant document(s).

Evaluation:
[Link].1 Has there been any evaluation of the efficacy of this law or policy? Please provide the link to any relevant
document(s).

12 The digital gender gap refers to the differences in access and use of digital technologies and the internet between genders. Please visit
[Link] and please provide the ITU Internet Gender Gap if available. Otherwise, please provide the Online
Internet Gender Gap (which is based on the Facebook Gender Gap Index, see here for more details: [Link]
indicators#internet).
13 [Link]
14 [Link]
15 [Link]
16 [Link] Science performance specifically measures the scientific literacy of
a 15 year-old in the use of scientific knowledge to identify questions, acquire new knowledge, explain scientific phenomena, and draw evidence-
based conclusions about science-related issues.
17 See National female e-inclusion policies, in the EIU Inclusive Internet Index: [Link]

14
 QUALITATIVE [Link] Has your country enacted any law or policy to reduce the digital socioeconomic or rural/urban
gap18? If not, is such a law or policy in the process of being adopted? Please elaborate and provide
the name and link to any relevant document(s).

Evaluation:
[Link].1 Has there been any evaluation of the efficacy of this law or policy? Please provide the link to any relevant
document(s).

[Link] Has your country enacted any law or policy related to enhancing diversity in the AI workforce? If not,
is such a law or policy in the process of being adopted? Please elaborate and provide the name and
link to any relevant document(s).

Evaluation:
[Link].1 Has there been any evaluation of the efficacy of this law or policy? Please elaborate and provide the link
to any relevant document(s).

[Link].2 Are tech companies required to publish diversity statistics?

[Link].3 Are affirmative action standards applied to improve diversity throughout the AI life cycle?

[Link].4 Are there outreach programmes to enhance diversity in STEM?

[Link].5 Are universities/PRO faculties required to publish diversity statistics?

[Link].6 Are government contractors required to adhere to diversity standards?

[Link] Is there online content and data available to train AI systems in all your country’s official languages19?

[Link] Is there online content and data available to train AI systems in your country’s indigenous
languages20?

3.2.2 Public engagement and trust

QUANTITATIVE [Link] Score of the country on the Online Services Index21

[Link] Score of the country on the E-Participation Index22

[Link] Trust in government websites and apps23

[Link] Trust in AI/trust in technology24

18 See Government e-inclusion strategy, in the EIU Inclusive Internet Index: [Link]
19 See Governance (G): Language Coverage and Data: [Link]
20 See Governance (G): Language Coverage and Data: [Link]
21 [Link]
22 [Link]
23 See Trust in Government Websites and Apps: [Link]
24 For instance, see [Link]
[Link]
science_00-06/, [Link]
Technology%20Final_10-[Link]

15
 3.2.3 Environmental and sustainability policies

QUALITATIVE [Link] Do you have in place any policy for addressing the impact of AI on the environment and on
sustainability? If not, is such a policy in the process of being adopted? Please elaborate and provide
the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of this policy been assessed? Please elaborate and provide the link to any relevant
document(s).

[Link].2 Is there an explicit mention of SDGs and/or ESGs?25

[Link].3 Is there a specific consideration for the impact of AI on land and water use?

[Link].4 Is environmental impact assessment mandatory before using AI in certain cases? Does this assessment
include a social component?

[Link].5 Is there a specific consideration for the environmental impact of AI demands on energy and its associated
carbon footprint?

[Link].6 Is there a specific consideration of the environmental impacts of the use cases that AI technology is
facilitating? (e.g. the effect of autonomous personal vehicles on transportation-related greenhouse gas
emissions, or using AI to increase fossil fuel exploration)26

3.2.4 Health and social well-being

QUALITATIVE [Link] Has your country adopted a digital health policy? Please elaborate and provide the name and/
or link to the relevant document. If not, is such a policy in the process of being adopted? Please
elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of the digital health policy (or equivalent) been assessed? Please elaborate and provide
the link to any relevant document(s).

[Link].2 Does your digital health policy (or equivalent) mention AI technologies?
[Link].3 Does your digital health policy (or equivalent) encompass physical and mental health?

[Link].4 Does your digital health policy (or equivalent) consider the impact of AI on children?

25 For definitions, see [Link]

26 For more on aligning AI with climate change mitigation, see [Link]
token=VNkvWalL1KVrxYte7VHfTtRgN0jAjWel9jnR3ZoTv0PFrrDwT9y68ah6oNg1ZvjBfcIb3Kbp_v5VihLBWDydbfc5jhcVPwJLve592DM4ofdyb9-
AdB37nZsc6NVbVb-fttOlElGgXuHDhQRUp82Rwc9buEoGydxnCgU609qUgUY%3D

16
3.2.5 Culture

QUALITATIVE [Link] Has your country implemented any policy regarding the use of AI for the preservation of cultural
heritage? If not, is there a policy on the preservation of cultural heritage that mentions the impact
of AI and digital technologies? Please elaborate and provide the name and link to any relevant
document(s).

Evaluation:
[Link].1 Has the efficacy of this policy been assessed? Please elaborate and provide the link to any relevant
document(s).

[Link].2 Does this policy allow community members to participate in the development or regulation of AI for the
preservation of cultural heritage?

[Link] Has your country implemented any policy regarding the use of AI for the preservation of minority
and indigenous languages? If not, is there a policy on the preservation of minority and indigenous
languages that mentions the impact of AI and digital technologies? Please elaborate and provide
the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of this policy been assessed? Please elaborate and provide the link to any relevant
document(s).

[Link].1 Does this policy allow community members to participate in the development or regulation of AI for the
preservation of minority and indigenous languages?

17
 4. THE SCIENTIFIC/EDUCATIONAL DIMENSION

4.1 Background
This dimension aims to assess the level of research and development related to AI in a country, including the number of AI-
related publications and patents, and the number of AI researchers and engineers involved in R&D. It will also look into AI ethics
research, e.g., reflecting the number of publications focusing on the ethics of AI. The educational dimension could include the
availability of education opportunities for students such as AI-related degree programmes, life-long education programmes
for AI developers and education opportunities for the general public. This aspect will examine, for example, whether there are
dedicated programmes that familiarize the general public and provide them with technology-related skills that could be useful
for keeping up to date with new job requirements. Apart from the education opportunities the dimension will examine the
numbers of students (graduates), professionals and the general public in AI- or ICT-related disciplines. It could also include the
number of graduates in STEM, as an important precondition for AI development and deployment. Potentially, the dimension
could address the level of AI and e-skills in the population. Finally, the dimension will include the availability of AI ethics education
for students of both AI-related and non-AI-related degrees as well as AI ethics courses for professionals including managers,
developers or product designers.

4.2 Indicators for the scientific/educational dimension

4.2.1 Research and innovation

[Link] R&D expenditure

QUANTITATIVE Evaluation:
[Link].1 Gross expenditure on Research and Development (GERD) as a share of Gross Domestic Product (GDP)27
[Link].2 Gross expenditure on Research and Development (GERD) on natural sciences and engineering as a share
of Gross Domestic Product (GDP)28
[Link].3 Government Budget Allocations for R&D (GBARD)29
[Link].4 Does your government have an estimate of government funding for Research and Development in AI? If
so, please specify how much and how it is broken down.

[Link] Research output

QUANTITATIVE Evaluation:
[Link].1 Number of AI and AI-related publications per capita30
[Link].2 Number of citations for AI and AI-related publications per capita
[Link].3 Number of FAccT publications per capita31

27 GERD is total intramural expenditure on R&D performed in the national territory (source: OECD Frascati Manual, [Link]
[Link] )
28 If the data is not directly available as a share of GDP, the GERD figure can be divided by GDP (both available through the OECD website)
29 GBARD encompass all spending allocations met from sources of government revenue foreseen within the budget, such as taxation (full definition
and source: OECD Frascati Manual, [Link] )
30 [Link]
31 The Association for Computing Machinery Fairness, Accountability, and Transparency (ACM FAccT) is the computer science conference with a
cross-disciplinary focus that brings together researchers and practitioners interested in fairness, accountability, and transparency in socio-technical
systems. The publications presented at the conference are published as conference proceedings. See [Link]

18
[Link] Ethical AI research

QUANTITATIVE Evaluation:
[Link].1 Number of publications on AI ethics per capita
[Link].2 Number of AI ethics conferences organized in the country per year per capita
[Link].3 Number of research centres and/or departments dedicated to AI ethics per capita
[Link].4 Number of AI research centres and/or departments that also cover AI ethics per capita

[Link] AI talent

QUANTITATIVE Evaluation:
[Link].1 Number of AI researchers (computer scientists, data scientists, roboticists, AI ethics researchers) in
universities/PRO per capita

[Link].2 Number of Kaggle grandmasters per capita32

[Link] Innovation output

QUANTITATIVE Evaluation:
[Link].1 Number of AI patents granted per capita33

[Link].2 Number of commits to AI repositories on GitHub per capita34

4.2.2 Education
[Link] Education strategy

QUALITATIVE [Link].1 Does your country have any laws or policies to integrate AI or other digital tools into the education
system? If not, is such a law or policy in the process of being adopted? Please elaborate and provide
the name and link to any relevant document(s).

Evaluation:
[Link].1.1 Has there been any evaluation of the efficacy of this law or policy? Please elaborate and provide the link
to any relevant document(s).

[Link].2 Does your country have any laws or policies on how educators/professors should be trained to teach
about AI/technology ethics? If not, is such a law or policy in the process of being adopted? Please
elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].2.1 Has there been any evaluation of the efficacy of this law or policy? Please elaborate and provide the link
to any relevant document(s).

32 [Link]
33 [Link]
34 [Link]

19
 [Link] Education infrastructure

QUANTITATIVE [Link].1 Proportions of primary, lower secondary and secondary schools with access to internet for
pedagogical purposes35
[Link].2 Proportions of primary, lower secondary and secondary schools with access to computers for
pedagogical purposes36

[Link] Curriculum content

QUANTITATIVE [Link].1 Number of tertiary education programmes dedicated to AI, machine learning or data science
per capita
[Link].2 Number of tertiary education programmes which offer one or more module in AI, machine learning
or data science per capita
[Link].3 Number of tertiary education programmes which offer one or more module in digital anthropology,
philosophy of technology, ethics of AI or related/similar disciplines per capita

QUALITATIVE [Link].4 Are there any educational programmes in your country that include both technical and ethical
aspects of AI (technical aspects might include coding, machine learning, statistics, data science etc.;
ethical aspects might include information ethics, philosophy of science and technology, privacy
concerns, social implication of technology, etc.)?

Evaluation:
[Link].4.1 In primary education? Here you may include courses to familiarize students with programming or digital
resilience (e.g., online safety, screen time, digital literacy).
[Link].4.2 In secondary education?
[Link].4.3 In universities and PROs?
[Link].4.4 In continuous education, vocational schools and technical/professional training institutes?

[Link] Educational attainment

QUANTITATIVE [Link].1 Percentage of STEM graduates in tertiary education37

[Link].2 Percentage of ICT graduates in tertiary education38
[Link].3 Number of graduates in data science, machine learning or robotics courses per capita
[Link].4 Number of AI-related PhDs per capita
[Link].5 Number of AI-related post-doctoral students per capita
[Link].6 Coursera Global Skills Report data science ranking39

[Link] Public access to AI education

QUALITATIVE [Link].1 Are there technical AI courses aimed at the general population?

Evaluation:
[Link].1.1 If yes, are they free and available in multiple languages?

[Link].2 Are there courses or modules on AI ethics aimed at the general population?

Evaluation:
[Link].2.1 If yes, are they free and available in multiple languages?

35 [Link] (specifically, in the “long format” section, under 4.a and 4.a.1)
36 [Link] (specifically, in the “long format” section, under 4.a and 4.a.1)
37 STEM stands for Science, Technology, Engineering, and Mathematics. For your country’s data, please visit [Link]
38 ICT stands for Information and Communication Technology(ies). For your country’s data, please visit [Link]
39 [Link]
20
5. THE ECONOMIC DIMENSION

5.1 Background
This dimension aims to address the size and strength of the supply side of the AI ecosystem in the country which is important
for the ability to develop AI solutions reflecting the particular needs and conditions of the given country and its population. It
will look at the size of the technology sector, including the number of companies which develop or deploy AI systems and their
employees. It will also address the amount of public and private investment in the field of AI. It will help track the growth of the
AI sector. Enhancing capacity in this regard would be related to the ability to support development of the AI ecosystem in the
country, including the attractiveness for investment in AI technology as well as in human talent.

5.2 Indicators for the economic dimension

5.2.1 Labour markets

QUANTITATIVE [Link] Share of job vacancies posted requiring AI-related skills (online job vacancies ideally)40

[Link] Share of current employees working as data scientists

[Link] Relative AI skill penetration41

[Link] AI talent concentration42

QUALITATIVE [Link] Does your country have a strategy to respond to AI impact on the labour market? This includes
issues such as re-skilling of workers affected by automation, upskilling of workers to take advantage
of opportunities presented by AI, and considering the soft skills advantages and complementarity
of human skills relative to AI systems. If not, is such a strategy in the process of being adopted?
Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of this strategy been assessed? Please provide the link to any relevant document(s).

5.2.2 Intermediate consumption

QUANTITATIVE [Link] How much do companies spend on AI services (including software as a service) as a share of
intermediate consumption (intermediate consumption of SIC 62)?43

Evaluation:
[Link].1 Do these AI services tend to be domestically produced or imported?

5.2.3 Investments and output

QUANTITATIVE [Link] Business Enterprise expenditure on R&D in the Computer programming, consultancy and related
activity sector per capita44

[Link] GDP for SIC code 62.0 (Computer programming, consultancy and related activities) per capita

[Link] High-tech exports as a share of trade45

40 [Link]
41 Defined as the prevalence of AI-related skills across occupations in the country, see: [Link]
42 Defined as the number of individuals with AI-related skills/in AI-related occupations on LinkedIn in the country, vis-a-vis the total number
LinkedIn members in the country: [Link]
43 Intermediate consumption consists of goods and services transformed or used up by the production process. Standard Industrial Classification
(SIC) 62 refers to Security and Commodity Brokers, Dealers, Exchanges, and Services.
44 Total intramural R&D expenditure performed by the Business Enterprise sector (source: Frascati manual)
45 [Link]

21
 6. THE TECHNICAL AND INFRASTRUCTURAL DIMENSION

6.1 Background
The technical and infrastructural dimension reflects the idea that without the relevant infrastructure, AI development and the
implementation of AI-based solutions cannot be scaled up throughout the country. Therefore, this dimension aims to assess the
level of ICT and related technical infrastructure in place. Among other things, the dimension will assess internet connectivity and
access, availability of data centres, cloud computing capabilities, and supercomputers. Given the critical importance of data for AI
technologies, another aspect within this dimension relates to the availability of high-quality data and practices for ensuring data
is representative. It should be noted that many indicators under this dimension are already measured by different indices, in the
Readiness Methodology we will pre-fill the answers and give countries the opportunity to update them if needed.

6.2 Indicators for the technical dimension

6.2.1 Infrastructure and connectivity

QUANTITATIVE [Link] Share of the population with a mobile telephone subscription46

[Link] Share of the population with a fixed broadband telephone subscription47

[Link] Share of the population with active mobile broadband subscription48

[Link] Average international bandwidth49

[Link] Average fixed broadband download speed50

[Link] Share of the population using the internet51

[Link] Share of population covered by at least a 3G mobile network52

[Link] Share of population with access to electricity53

[Link] Gender gap in internet access54

[Link] Gender gap in mobile access55

[Link] Rural/urban gap in internet access (households)56

46 [Link]
47 [Link]
48 [Link]
49 International bandwidth refers to the total used capacity of international bandwidth; in megabits per second (Mbit/s). It is measured as the
sum of used capacity of all Internet exchanges (locations where Internet traffic is exchanged) offering international bandwidth. If capacity is
asymmetric (i.e. more incoming (downlink) than outgoing (uplink) capacity); then the incoming (downlink) capacity should be provided. See:
[Link]
50 [Link]
51 [Link]
52 [Link]
53 [Link]
54 [Link]
55 [Link]
56 [Link]

22
6.2.2 Applied standards

QUALITATIVE [Link] Is your country involved in standardization (both technical and ethical) of AI and digital technologies?
(ISO/IEC, IEEE7000)57?

[Link].1 If not, have you applied to be part of this process?

6.2.3 Computing capabilities

QUANTITATIVE [Link] Number of data centres in the country per capita58

[Link] Distance to the closest data centre59 in km

[Link] Colocation data centres60 per million population

QUALITATIVE [Link] Does your country have a policy for AI-driven cloud computing? If not, is such a policy in the process
of being adopted? Please elaborate and provide the name and link to any relevant document(s).

Evaluation:
[Link].1 Has the efficacy of such policy been assessed before? Please provide the link to any relevant document(s).

57 The International Organization for Standardization (ISO) and the International Electronic Commission (IEC) Joint Technical Committee (JTC 1)
for information technology, is a consensus-based, voluntary international standards group. Over 2000 experts from 163 countries develop
mutually beneficial guidelines that enhance global trade while protecting intellectual property. Here is a list of relevant subcommittees and
working groups handling both technical and ethical standardization in AI: Subcommittee (SC) 40 on Governance and Service Management,
SC 41 on Internet of Things and Digital Twin, SC 42 on Artificial Intelligence, Big Data Programme, SC 42 on Artificial Intelligence, Subgroup (SG) 2
Trustworthiness Study Group.
The Institute of Electrical and Electronics Engineers (IEEE) is the world’s largest technical professional organization dedicated to advancing
technology for humanity. Several Standards Working Groups are currently active under the P7000 standards series on socio-technical
issues related to AI: IEEE P7003: Algorithmic Bias Considerations, IEEE P7004: Standard on Child and Student Data Governance, IEEE P7004.1:
Recommended Practices for Virtual and Classroom Security, Privacy and Data Governance, IEEE P7008: Standards for Ethically Driven Nudging for
Robotic, Intelligent and Autonomous Systems, IEEE P7009: Standard for Fail-safe Design of Autonomous and Semi-Autonomous Systems, IEEE
P7010.1: Recommended Practice for Environmental Social Governance (ESG) and Social Development Goal (SDG) Action Implementation and
Advancing Corporate Social Responsibility, IEEE P7011: Standard for the Process of Identifying and Rating the Trustworthiness of News Sources,
IEEE P7012: Standard for Machine Readable Personal Privacy Terms, IEEE P7014: Standards for Ethical Consideration in Emulated Empathy in
Autonomous and Intelligent Systems, IEEE P7015: Standard for Data and AI Literacy, Skills, and Readiness. For more information, see
[Link] and [Link]
58 A data centre is a facility that centralizes an organization’s shared IT operations and equipment for the purposes of storing, processing,
and disseminating data and applications. Most modern data centre infrastructures are not only physical: virtualized infrastructure supports
applications and workloads across multi-cloud environments (source: [Link]
Please note that you may consider Public Research Institutions and Universities here, as they sometimes host data centres that are also
computing centres (for instance, the Centre Informatique National de l’Enseignement Supérieur in Montpellier, France).
See [Link]
59 [Link]
60 Type of data centre (see note 52 for definition) where equipment, space, and bandwidth are available for rental to retail customers.
See: [Link]

23
 6.2.4 Statistical performance

QUANTITATIVE [Link] Statistical performance indicators61

[Link].1 Score on data products62

[Link].2 Score on data sources63

[Link].3 Score on data infrastructure64

QUALITATIVE [Link] Are there any laws or policies providing a comprehensive framework for consistent data
management and publication?65

Evaluation:
[Link].1 Has the efficacy of such policy been assessed before? Please provide the link to any relevant document(s).

[Link].2 Are there clearly documented quality control processes for government data?66

61 [Link]
62 Defined as the availability of data for the 17 SDGs (social, economic, environmental and institutional statistics coming from the UN SDG
database). The products signal whether countries are able to produce indicators related to the 17 Sustainable Development Goals.
63 Defined as the availability of recent censuses and surveys covering broad areas (population & Housing census, Agriculture census, etc.), the
availability of Civil Registration and Vital Statistics (CRVS), and the availability of geospatial data.
64 A mature statistical system has well-developed data infrastructure, evidenced by legislation and governance (Legislation Indicator based on
PARIS21 indicators on SDG 17.18.2), as well as standards and methods of data compilation (system of national accounts in use, National Accounts
base year, Classification of national industry, CPI base year, Classification of household consumption, Classification of status of employment,
Central government accounting status, Compilation of government finance statistics, Compilation of monetary and financial statistics, Business
process)
65 See Governance (G): Data Management: [Link]
66 See Governance (G): Data Management: [Link]

24
Annex
 Annex

Country level implementation outline

As mentioned above, the basic workplan for conducting the Readiness Assessment is built on key elements and offers flexibility
to be adapted to the unique circumstances in each Member State, based on the consultation with the relevant field offices and
the national stakeholders. Thus, the plan outlined below meant to serve as example only, and to illustrate the different stages that
will be part of the project. Please note that the intended readers of this annex are UNESCO field offices that will be tasked with
the delivery of the project in each country, in collaboration with other stakeholders.

1. Beneficiary countries identified based on the consultation with the field offices, HQ and/ or the requests received from the
Member States.

2. In each country, UNESCO will pursue work on two parallel tracks:

Securing and maintaining high-level political support for conducting the assessment, to allow for an unimpeded engagement
of the team working on the ground with the relevant government agencies.

Launching a National Stakeholder Team to lead on the Readiness Assessment. Ideally, the team would include:

• SHS Programme Specialist from a regional UNESCO Field Office,

• Representative from the National Commission
• Focal point from the ministry with AI/Digital governance portfolio
• Representatives of other ministries (such as the ministry of education science and technology, the ICT ministry, and the
ministry of research),
• Representatives of the academic community (if possible, scholars working on AI/ digital transformation from the faculty
of law, public policy, economy, and computer science),
• Representatives of civil society
• Representative of the private sector

3. A local consultant/ implementing partner could be hired to coordinate the team and consolidate the output of different
members. Depending on the regional and national context, and the available funding, this could be one person responsible for
several countries, or one consultant in each of the beneficiary countries. The team will be supported by a member of the UNESCO
Secretariat at HQ, and as needed also by an independent international expert in AI ethics, who will assist the team. HQ is currently
setting up the AI Ethics without Borders network, which will also be deployed to assist with conducting readiness assessment or
specific aspects of it.

4. A national launch event will be organized in each country (alternatively, based on the preferences of a field office and
budget availability, one regional event could be organized with the teams from the beneficiary countries participating). This
event will feature:

High level political segment demonstrating government’s buy in and support for the project.

The launch of the national stakeholder team.

Training delivered for the country team by UNESCO on the implementation of the Recommendation on the ethics of AI, focusing
on the Readiness Assessment methodology.

Discussion and agreement on the concrete action plan with milestones and deliverables for each country.

5. Mid-term workshop could be held for the National Stakeholder Team to take stock of the process, discuss the missing
elements and brainstorm solutions for how to address them. At this stage, contact with the high-level officials supporting the
process might be needed in order to facilitate access to types of data that was not easy to gather.

6. Concluding workshop will be held towards the end of the assessment, to present a draft country report and to receive
feedback from a broader set of stakeholders on the way forward. The Ethical Impact Assessment tool will also be presented

26
during the concluding workshop, based on the results of the readiness assessment, for the National Stakeholder Team to discuss
its adoption in the specific context.

7. Upon the finalization of the readiness report, it will be disseminated widely at the national level among the main stakeholders
and made public on the UNESCO observatory on the Ethics of AI.

As part of the final report, a roadmap will be designed in consultation with the beneficiary countries, highlighting the conclusions
and suggesting a path forward that reflects the priorities of the country in terms of the specific institutions that need to be built
or enhanced to implement the Recommendation.

This process is intended to be cyclical in nature—Member States will have the possibility to contact UNESCO to re-evaluate
their readiness to implement AI ethically and to track their progress on the roadmap over time. UNESCO will keep a record of all
documents produced during the initial readiness assessment and build on these documents.

27
Social and Human Sciences Sector
7, place de Fontenoy
75352 Paris 07 SP France
ai-ethics@[Link]
[Link]/Ethics-of-AI
Supported by

Follow us
@UNESCO #AI #HumanAI