Sunteți pe pagina 1din 35

A Roadmap for AI for

Development in Africa
Supported by
Government AI Readiness Index
Cluster Indicator
Governance Data protection/privacy laws - yes/no
National AI strategy - yes/no/pending
Infrastructure and data Data availability
Government procurement of advanced technology
Data/AI capability (in government)
Skills and education Technology skills
Private sector innovation capability

Number of AI startups
Government and public services Digital public services
Effectiveness of government
Importance of IT to government’s vision of
the future
Goals for
workshop
• scope out the African ML/AI
landscape
• Create an African AI research
roadmap
• Support the development of
cross-continent cooperation
What have we
discussed?
Policy and Regulation: Shape ethical and rights based AI

Capacity: Build local skills in AI

Applications: Support locally driven applications for the public


interest
Governance
Key Needs

Designing regulatory frameworks that both stimulate AI innovation, while minimizing harms (to
civil liberties) – the latter should not stifle the former - these frameworks don’t exist yet
African context: Need regulatory frameworks that are designed for the African context, not simply
importing western-designed frameworks
Need for regulatory learning - ensure regulation is learning from advancements in AI, so that
regulation doesn’t stifle AI advancements wit policy rapid responses
Trust: Need for greater transparency and accountability of algorithms for greater trust. How do we
do that?
Quantification of harm: Need research to identify and quantify potential harms to be able to gauge
whether it is a an acceptable level of risk – as well research to quantify the benefits of AI
(Do we need pre-emptive regulation for pre-emptive harm?)
Existing Initiatives

• PASET: Partnership for skills in applied science engineering and technology (AU) – 4th
industrial revolution forum – Kigali May
• AI and blockchain task force in Kenya
• RIA: Research ICT Africa – policy papers
• PRIDA – Policy and Regulation Initiative for Digital Africa (AU-EU)
• Open AIR: AI innovation
• RightsCon – Tunis – algorithmic transparency
• DRIF: Digital Rights inclusion Forum DRIF’19 – Lagos yearly
• AUDA – NEPAD: flagship projects – AI
• ACFTA: Africa Continental Free Trade Area
5 Year Targeted Outcomes
• 30 African countries to have AI specific policies

• To have 15 ratifications and 30 signatures to


African Data protection and Cybercrimes
Convention

• To have quality AI policies that include:


• Multi-stakeholder and evidence based input
• Systematic inclusion of normative framework
(laws) into system protocols.
• AI regulation that is ethical and rights based
• Context specific regulation for Africa
Policy Research Capacity Goal

• To have 5 regional research centers that have the long term vision of being
centers of excellence in Africa
• Towards a center of excellence in each country that are responsive to AI policies and
windows of opportunities

• To have 500 researchers which includes:


• Policy researchers that understand AI
• AI researchers who understand policy
• Majority works for the public interest
Effective Global Policy Engagement

• Participation:
• Guarantee a minimum number of Africans to be involved in setting the policy and
policy research agenda globally;
• Ensure African researchers are meaningfully involved in ALL global AI conversations;
• Research Funding support:
• just in time and long term research questions
• Support for effective research communications
• Activities:
• Peer mentorship for quality research;
• Access to journals and legislation portals;
• Online and event-based knowledge sharing facility.
Short-term steps

• Online network sharing


• DRIF April 23-5th – Workshop on Algorithmic Transparency – Arthur
Gwagwa
• Rightscon meeting in June to ID next steps
• AI for Good Summit 28-31 May 2019 - Geneva, Switzerland
• IGF meeting – Berlin November – further progress with new
stakeholders (12th April workshop deadline)
Capacity
The Philosophy

Capacitating AI Strength through


Community
Key Needs

• Context [Environment, Utility]


• Need to understand what the local needs are
• Why is it useful to have AI/ML [Capacity <> Funding]
• Educational Agility and Clarity
• Adapt to the local need [Research, Startups etc.]
• Transdisciplinary
• Explain and demystify AI/ML for the public. Better entrance pipelines.
• Synergy
• We need to understand what is going on around the continent
• Sharing resources [Abundance in funding, scarcity in opportunity]
• Funding
• Mentorship
• Collaboration
Existing Initiatives

Existing initiatives
New ideas
• Regional Community Capacity Development Coordination
• Southern African Region
• East African Region
• West African Region
• North African Region
• Integration and sharing across regions. Nodes and SuperNodes.
• Graduate 400 African PhDs in ML/AI/DS in next 5 years?
• Have a research center in every African country?
• Strength at all Levels = startups / education / research /
training / teaching / industry / meetups = World-Class!!
• Break Dichotomy between Research and Application
• Collaboration with IP protections
Key Activities and Outcomes
Activity Outcome(s)
Study on best practice and defining AI for Africa Informing strategic direction [Look at successful Afrrican countries]
Communicating AI to the Public Understanding by the public [general, secondary schools, etc.]

AI Talent Pipeline

AI Centers of Excellence [People, not places] Skilled pipeline for the continent.
Startup growth
Research Cohorts [MS, PhD, Faculty] Industry Growth
Research outputs and impact
AI Engineering Cohort - Compressed MS Better understanding of business/society/science value
Effective Mentoring
AI Engineering+ [Industry Training]
Open funding for capacity building [Strategic] Localised capacity building nudging collaboration across academia,
government, industry and startups.

Public Policy and Strategies [Government etc.] Governments enter the game, shape strategy and deliver.

Multi-disciplinary Co-Creation Cross cutting innovation

Challenges & Datasets Innovation through Friendly Competition

Synergy Coordination Understanding who and where: Industry, Startups, Research


Applications
Key Needs
• Peer Group For AI Practitioners
Providing amplification to private sector needs and strengths
• Access To Funding
Capital for infrastructure, PM Fit, Growth Stage ventures is rare, government funding for risky
projects
• Africa's Fragmented Market
Challenge to getting critical mass of Customers for AI innovations
• Limited access to information
Limited access to academic research, open science, methods, data
• Limited infrastructure
Electricity, computing speeds, cloud space
• Response to SDGs
Private Sector and Innovations needs to respond to needs in Africa now and in future
Existing Initiatives

• H3Africa – Human Heredity and Health in Africa


• Global Pulse: UBOS
• Data Science Africa
• Women in data science and machine learning: agriculture
• Efficient agriculture: connecting use of drones with phenotype/genotype data
• Research to create personalize medicine in Western Cape Town
• Social media to prevent spread of dengue – Sierra Leone, Burkina Faso
• Nigeria University - AI for diagnosis, clinical decision support, digital, assessment
of diseases
• Kenya indigenous poultry consortium - Social media mining for students to
communicate; Disease outbreak of indigenous chicken
Existing Initiatives

• BBC/Edinburg - Translation of African languages into English


• University of Cape Town - Information retrieval, cultural heritage preservation
work (historical languages….), machine learning in educational technology; also
• Many research groups - Data Science Africa
• Element AI - Satellite imagery to identify signs of conflict, impact of natural
disasters in communities, climate change,
• Google Launchpad Studio
• Cortex Logic VC
• Africa AI show
• Facebook
• Microsoft HeadStart
• IBM Research Labs
New ideas

• Collaboration with non-AI communities


• Challenges to scale up of best practices
• Buying from local actors
• Funding
• Policy, legislative and regulatory frameworks
• Academic Curriculum Revision
Points for Reflection

• How might we incentivize private sector to work on AI


with strong development outcomes (i.e.AI4D)?

• How might we support & stimulate private sector and


academic research partnership?

• How might we develop a funding mechanisms


(including start-up & venture capital) to support local
development/use of AI?
Key Outcomes (5 year target)

• Strong networks that includes 10-20 leading African-based companies substantially


contribute to key areas of the African development agenda
• Networks of telecoms, Large tech companies, universities, research centres (ILRI, ICIPE,
KEMRI, etc.), Public Sector Institutions (Example: MTN Element AI, and users in Africa)
• Focus on skill development and development of local solutions

• Funding to collaboration on AI
• Investments in AI to advance collaboration between private sector, universities and public-interest
institutions. (USD 1 Billion target over the next 5 years?)
• Engage VCs, Angels, Unrestricted grants to support startups for local use cases in AI for
development

• Global resources available for African entrepreneurs


• Organising AI conferences in Africa with diverse representation
• Networking platforms with global players interested in building resources
• Facilitate tech transfer to local actors
Key Outcomes (short term)
• Linkages with existing networks of practitioners
• Support and learn from collaborative efforts of African organisations
focused on AI4D
• Explore quick wins:
• Pilots between academia and sector domain experts
• Embed AI fellows in NGOs etc.

• Research
• Identify viable AI4DEV solutions (i.e. fast implementation, massive
scale)
• Need to understand AI innovation cycle in Africa, including academy
participation
Reaching the Marginalised and the Poor

• Identify and work with development communities e.g. Invite


development orgs to Deep Learning Indaba or other events to
connect
• Include diverse data sources for AI Innovation e.g. platform on big
data from CGIAR, Telco data
• Integrate AI into development projects from an early phase
• Engage with standard setting organizations to ensure innovations
scalable across use cases
• Engage local leadership during problem solving and product
development cycle
Addressing Africa's Diversity

• Encourage dissemination of information.


• AI Wiki, for example

• Recognize the inhomogeneity of Africa:


• Specific conversations with Francophone and Portuguese-speaking Africa
• Engage different countries to help with contextually-relevant groupings
• Create platforms for engagement of different groupings
• Engage regional bodies for identifying goals
The work is just starting...
In the next months, we will continue to engage
other stakeholders in shaping these plans

Website: ai4d.ai
https://ai4d.ai/event/ssa-network/
Twitter: @ai4dev #ai4dev
idrc.ca/ai4d
Focal points
Policy and Regulation Capacity building Applications

Alex Comninos Ciira Maina Leonida Mutuku


Researcher Senior Lecturer CEO / Researcher
Research ICT Africa Dedan Kimathi University of Intelipro Limited / LDRI
https://comninos.org Technology https://www.developlocal.org/
www.ciirawamaina.com
Dr. Robert Muthuri Kathleen Siminyu
Research Fellow Vukosi Marivate Head of Data Science
ICT at the Centre for IP and IT (CIPIT) ABSA UP Chair of Data Science Africa's Talking
Strathmore School of Law Dept. Computer Science, Univ. of Pretoria https://africastalking.com/
www.vima.co.za
Olubayo Adekanmbi
Tejumade Afonja
Chief Transformation Officer/
AI Engineer
Convener MTN Nigeria/
AI Saturdays Lagos /
Data Science Nigeria
InstaDeephttps://aisaturdayslagos.github.io/
www.datasciencenigeria.org
Aisha Researcher IBM Research Africa John Shawe-Taylor Professor and Head of Department UCL
Alex Comninos Researcher Research ICT Africa
Juliet K. Mutahi Software Engineer IBM Research | Africa
Alexandra Birch Research Fellow University of Edinburgh
Alfred Ongere Founder and team lead Ai Kenya Program Manager, BioInnovate
AnnaMaria Oltorp IDRC Julius Ecuru Africa icipe
Arthur Ernest Gwagwa Research Associate CIPIT: Strathmore Law School Kathleen Siminyu Head of Data Science Africa's Talking
Kommy Chief Scientist IBM Research - Africa
Senior Lecturer / Principal Kris Sankaran Postdoc Mila
Benjamin Rosman Researcher University of the Witwatersrand / CSIR
Langa Khumalo Professor University of KwaZulu-Natal
Betty Murithi Strathmore University Sida Research Programme
Bitange Ndemo Associate Peofessor University of Nairobi Latifa Hamisi Mbelwa Coordinator University of Dar es Salaam
Bolanle F. Oladejo Senior Lecturer University of Ibadan Laurent Elder IDRC
Leonida Mutuku CEO Intelipro Limited
Brian Muhia Machine Learning Engineer UTU Technologies MALO Sadouanouan Lecturer and researcher Nazi BONI University
Maria Faslli University of Essex
Charity Wayua Senior Research Manager IBM
Dedan Kimathi University of Melisa Achoko Allela Assistant Lecturer Technical University of Kenya
Ciira Maina Senior Lecturer Technology Director Research, Innovation and
Cynthia M. Mulenga Product Manager BongoHive Moses Thiga Outreach Kabarak University
Daniele Orner Chief Scientist Brave Muthoni Wanyoike Team Lead Instadeep
Naser Faruqui IDRC
Davor Orlic COO Knowledge 4 All Foundation
Neil Butcher Director Neil Butcher & Associates
Delmiro Fernandez-Reyes Dr. University College London
Chief Transformation Officer/
Dept of Comp. Science, University of Olubayo Adekanmbi Convener MTN Nigeria/Data Science Nigeria
Associate Cape Town/Centre for Artificial Olubunmi Banjo Managing Director Kuvora
Deshen Moodley Professor/Deputy Director Intelligence
Dr Nick Bradshaw CEO AI Media PTY Ltd Manager Pulse Lab Kampala (UN
Swedish University of Agricultural Paula Hidalgo-Sanchis Global Pulse) UN Global Pulse
Erik Bongcam-Rudloff Professor of Bioinformatics Sciences Jomo Kenyatta University of Agriculture and
Eva Ohlsson IDRC Philip Apodo Oyier Assistant Lecturer Technology (JKUAT)
Fernando Perini IDRC
Geoff Hill Senior Policy Manager UKRI Prateek Sibal Specialist Consultant - AI UNESCO
Grace Mutungu Associate Kenya ICT Action Network Robert Peeters Netherlands Embassy in Keny

Head of Department and Department of Computer Science, Sarah Kiden Technologist and Researcher Research ICT Africa
Hussein SULEMAN Associate Professor School of IT, University of Cape Town Sherif Elsayed-Ali Director of Partnerships Element AI
Ivy Barley CEO Developers in Vogue T. Idriss TINTO Coordinator CAFDO
Jeph Acheampong Founder Blossom Academy Tejumade Afonja AI Engineer AI Saturdays Lagos / InstaDeep
Program officer, Research Tonya Nyakeya THINKLab & Ecosystem Lead IBM Research - Africa
Johan Hellstrom and Innovation Embassy of Sweden, Dar es Salaam
Ulrich Paquet Staff Reseach Scientist DeepMind / Deep Learning Indaba
University Relations & D-
John Matogo NA Leader for East Africa IBM Vukosi Marivate ABSA UP Chair of Data Science Dept. Computer Science, Univ. of Pretoria
Organizers

• The International Development Research Centre (IDRC) funds research in developing countries to promote
growth, reduce poverty, and drive large-scale positive change. IDRC seeks to co-design a AI for Development
initiative, with a focus on research and capacity-building to develop AI applications that are inclusive, ethical,
and rights-based. Through this initiative, funder partners will pool and leverage resources to help shape the
rollout of AI in the global south, mainly through strengthening Southern institutions.
• The Knowledge 4 All Foundation is the only European machine learning focused not-for-profit and an
advocate of AI applications for reaching the UNs Sustainable Development Goals (SDG), especially SDG4. It
originated from the PASCAL Network which was an EU funded Network of Excellence comprising some 1000
machine learning, statistical and optimisation researchers. It has de facto leading and building the
international machine learning community, through its activities that culminated in today’s AI hype,
including supporting the NIPS conference.
• The Swedish International Development Cooperation Agency (SIDA) is a government agency of the
Swedish Ministry for Foreign Affairs. Sida is responsible for organization of the bulk of Sweden's official
development assistance to developing countries.
• Strathmore University is a chartered university based in Nairobi, Kenya. Strathmore College was started in
1961, as the first multi-racial, multi-religious Advanced-level Sixth Form College offering science and arts
subjects, by a group of professionals who formed a charitable educational trust.

S-ar putea să vă placă și