Download full agenda with speakers (latest update: 23 October)

Thursday 26 October: AI developments and applications

Session 1. State of AI research

9:45-11:00

This session will introduce the distinctive characteristics of artificial intelligence and machine learning. It will provide an overview of milestones to date in AI development and of expected future milestones. For example, standalone AI is expected to evolve towards networks in which AIs communicate and interact. The session will describe some of the focus areas of both private and public investment in AI research. It will also introduce the link between AI and robotics.

Session 2. AI applications and case studies

11:30-12:45

This session will illustrate how AI is being applied to make better decisions, reduce costs and improve productivity in a variety of domains. In environmental applications, AI can find complex causalities among environmental variables and optimise resource use. In health, AI can help detect conditions early, deliver preventive services and discover new treatments. In transportation, autonomous driving and optimised traffic routes can facilitate efficient travel and save lives. In security, AI can help identify & combat real-world and cyber threats. More broadly, AI is finding valuable application wherever intelligence is deployed, including unexpected areas such as arts & culture and services of all kinds.  

Session 3. Close-up on AI in space applications

13:45-15:00

In the space industry, new and improved satellite data and signals combined with AI are powering innovative products and services in sectors such as finance, agriculture, land use, and disaster management. In this session, innovative start-ups, space agencies and space administrations will illustrate the growing connections between AI and satellites.

 

Session 4. Enhancing discovery: The role of AI in science

15:00-16:30

AI promises to improve research productivity at a time when ideas are becoming harder to find, pressure on public research budgets is increasing, and global challenges require scientific breakthroughs. This session will explore the opportunities and challenges of applying AI in science, by examining:

  • current and emerging uses of AI and machine learning in science
  • limitations in using AI in science
  • the opportunities that machine learning presents to increase research productivity
  • challenges posed for researchers
  • issues raised for institutions (including science education), research sponsors and policymakers.

 

Session 5. AI policy landscape

17:00-18:30

This session will provide an overview of the AI policy landscape, covering national and non-governmental initiatives. Governments and a range of institutions are holding national consultations on opportunities and challenges associated with AI and developing national AI-related strategies that often intersect with robotics. Stakeholders from the private sector, research communities, civil society and trade unions are actively examining AI-related issues. For example, some of the most active AI companies have engaged in a Partnership on Artificial Intelligence to benefit people and society; the IEEE Standards Association has created the “Global Initiative for Ethical Considerations in the Design of Autonomous Systems”; and AI-focused foundations and institutions are developing. The discussion will begin to identify commonalities and differences between value sets to help technologists design AI systems compatible with societal norms.

Friday 27 October: Public policy considerations raised by AI

Session 6. Employment and skills

9:30-11:00

AI will generate new jobs and occupations. But it is also expected to create challenges for employment by replacing human labour and eliminating some occupations, not just in low-skilled work but also in highly-skilled professions such as radiology (e.g. with advances in image and pattern recognition). This session will discuss the types and numbers of jobs vulnerable to replacement by AI. It will look into risks of increasing unemployment and earnings inequality, new needs for skills development and training, and the role of governments and public policy. 

Session 7. Safety, responsibility and liability

11:30-12:30

AI-driven automated decision-making raises questions of safety, responsibility and liability, for example when accidents involve autonomous vehicles. This session will discuss responsibility for AI-powered decisions and the respective roles of actors such as AI software developer, hardware provider, owner, driver/passenger etc. It will examine the applicability of existing mechanisms such as insurance to deal with uncertain risks and whether legal clarification is called for.

Session 8. Privacy and security

14:00-15:00

Data generated by users, consumers and businesses train machine learning algorithms, which require vast amounts of data to recognise patterns efficiently. This session will examine how machine learning approaches combined with ever-increasing amounts of personal data affect the concepts of privacy and consent and what actions are needed to help guide AI developments towards social goals. The session will also discuss new security risks linked to misuse of AI through negligence or malice, for example of malware abusing an AI network system or an autonomous weapon.

Session 9. Transparency, oversight and ethics

15:00-16:00

This session will examine how to ensure that AI-powered decisions that impact people are fair, transparent and accountable while preserving legitimate commercial confidentiality and productivity. Issues are already manifest in critical areas such as determining priorities in hospital line-of-care, autonomous vehicle emergency responses, criminal risk-profiling and preventive policing, and access to credit and insurance. The session will also provide an opportunity to discuss:

  • public acceptance of AI
  • risks of algorithms amplifying social biases and discrimination
  • the growing difficulty of understanding AI algorithm decisions
  • the role of new technologies that can provide rationale for an algorithm’s decision
  • respective stakeholder roles
  • cost considerations for developing and implementing algorithmic accountability solutions at scale. 

Session 10. Wrap-up and next steps

16:30-18:00

This session will explore the AI governance landscape and the respective roles of industry self-regulation, policy interventions, multi-stakeholder co-operation, and international dialogue. It will consider the degree to which existing digital economy policy principles could help address new challenges presented by AI or need adapting and discuss the role of international co-operation.