introduction

Artificial intelligence (AI) is a hot topic right now. There is a huge amount of excitement and hype about the ways in which AI could reshape our lives and society in coming years. We are already seeing the early stages of many potentially transformative developments, from the introduction of autonomous vehicles on our roads and the use of AI-powered “robo-advisors” in financial services, to the application of machine learning in medical research to model complex molecules in unprecedented ways. Not to mention the use of algorithms that help deliver more effective and targeted public services.

At the same time, there is also growing awareness of the potential dangers of developing and implementing a technology of this power without proper care and oversight. Issues such as the danger of bias in algorithmic decision processes, the impact of automation on the future of the workplace, and the use of AI to undermine and distort democratic processes have become key parts of mainstream political and cultural discourse. As a result, the debate about AI ethics ─ and what may need to be done in terms of new legislation and regulation ─ has moved from the margins of academia to the highest corridors of power.

If AI has anything like the impact that many are predicting, it will have profound implications for civil society and the work of charities. This is likely to play out in three broad ways:

1 New ways of achieving a mission

Civil Society Organisations (CSOs) can harness AI to find new ways of delivering social and environmental good.

2 Impact on organisations

AI will make it possible to automate many services, create new models of governance and transform the wider operating environment. This will affect CSOs in the same way it does organisations from other sectors.

3 Creating new challenges

AI (just like any other technology) has the potential to create new problems and challenges for our society. These may be unintended negative consequences, or they may be the result of deliberately malicious uses of the technology by malign actors.

the challenge for civil society

Civil society can play a key role in addressing the risks posed by AI. And it is crucial that it does, because CSOs represent the most marginalised individuals and communities in society. Since these will be the people hit earliest and hardest by the negative consequences of technology, they need someone to speak up on their behalf.

The challenge for CSOs is therefore threefold: to understand the positive potential for harnessing AI to address social and environmental problems; to understand the impact of AI on organisations, industries and the workplace; and to understand the negative impacts on the communities they serve and on wider society.

To address these three challenges, it is vital that CSOs keep abreast of the issues and how they potentially relate to their work. It is also crucial that there are opportunities to work with the tech sector to explore the possibilities for using AI for social good. Finally, it is critically important that civil society has a seat at the table in the debate over AI ethics, as many CSOs will have unique and valuable insights to offer, and the cost of missing out on this could be huge.

aims of this paper

This Giving Thought discussion paper 'Machine made goods:Charities, Philanthropy & Artificial Intelligence' aims to:

  • Spark thought and debate about both the good and bad sides of AI and this technology’s potential in the context of civil society
  • Consider what needs to be done by civil society itself, by the private sector, and by government to ensure that the long-term impact of these technologies is primarily positive rather than negative.

This sits within CAF’s wider Future:Good project, in which we are looking more broadly at the impact of disruptive technologies on civil society, philanthropy and the work of charities.

why is ai such a big deal right now?

Artificial intelligence as a field of study has been around for a fairly long time. Its roots can be traced back at least as far as a series of papers by computing pioneer Alan Turing which paved the way for the concept of “machine learning” that sits at the heart of modern AI.

This is the idea that rather than attempting to create an AI system by specifying up front all the rules that will determine its operation, we instead mirror the way we actually develop as humans, by creating algorithms that have the ability to “learn” through a process of repeated trial and self-modification.

As an interesting historical side-note for those interested in the intersection of charity and AI, philanthropy actually played a key role in the early development of the technology. Between 1932 and 1955, the Rockefeller Foundation employed as its Director of Natural Sciences (or “Chief Philanthropoid” as he preferred to style himself) a man named Warren Weaver.  Weaver was a high-quality mathematician in his own right, but he also played a hugely important catalytic role in his time at Rockefeller by using the foundation’s grantmaking to support scientific research in a range of areas, including the nascent field of computer science. In 1956, Rockefeller foundation gave a grant of $7,500 to support a conference in Dartmouth that is now widely recognised as the birthplace of modern AI, and it was in fact in the grant application for this conference that the first recorded usage of the term “Artificial Intelligence” can be found.

Given this long history, why has AI all of a sudden come to an unprecedented level of prominence in recent years? There are a number of factors that can help explain this, including:

  1. New types of algorithms: there have been huge leaps forward in the field of machine learning (ML) in the last decade; most notably the development of a type of powerful new algorithm known as a “deep neural network”.
  2. Data explosion: ML algorithms require vast quantities of data on which they can be trained; and historically this was a major limiting factor. However, recent years have seen an exponential increase in the amount of data being generated in all areas of our lives, and this has provided the fuel for the rapid growth and development of ML.
  3. Processing power: The leap forward in software terms represented by the development of deep learning algorithms has only been possible because of an associated leap forward in hardware. In particular, the development of fast, cheap Graphical Processing Units (GPUs), originally driven by computer gaming market, has made the kind of processing required for ML affordable at scale.
  4. Investment: As the potential of AI has become clear (given the factors outlined above), it is no surprise that investors are piling in to the field and putting up huge sums of capital to fund new developments and applications. In particular, the Chinese government has started funding AI very aggressively, which has led many experts to suggest that we may be in the early stages of an “AI arms race”.