Robotic alms ai and the future of charitable giving notes
1. Rhodri Davies- Programme Director, Giving Thought at Charities Aid Foundation
Robotic Alms: AI and the Future of Charitable Giving
In broad terms, AI technologyis likely to affectcharities in four key
ways:
i) Creatingnew problemsthat charities willbe called upon to
address:AI,like many technologies,will have unintended negative
consequences,which charities will be relied upon to solve.
ii) Developing new ways of addressing existingproblems: AI
allows the analysis of data at an unprecedented scale and speed,
which could suggestcompletelynew ideas for solving social and
environmental problems.
iii) Offering new ways of working for existingcharitable
organisations: AI could transform things like regulation and financial
reporting, and could also radically alter the ways people are able to
give.
iv) Creating new governancestructures and operating models
for achievingsocialgood:AI could lead to ways of working which
augment or even replace traditional charitable organisations entirely.
Happy to discuss any of these further, but going to focus in on (iii),
and in particular the impact of AI on giving and fundraising.
What AI could revealaboutphilanthropicand fundraising
motivations
IN terms of funding from organisations with clear existing purpose
and expertise (i.e. grantmakers), AI could have major impact in
making system more efficient
CF: Geoff Mulgan Nesta blog:possibilityof using machine learning
to evaluate grant applications.
2. Rhodri Davies- Programme Director, Giving Thought at Charities Aid Foundation
Definitely a lot of merit in this, given how opaque grant application
and awarding process oftenis.
HOWEVER:I’m going to focus in on something I have written in
more detail about, namely impact of AI on philanthropy advice
Philanthropy advice is something that has been around for a long
time
I could give you some historical examples like 17th
century London
Merchant Thomas Firmin, or Charles Dickens
HOWEVER:largely the preserve of the wealthy, as it is a bespoke
service and only economicallyjustifiable if you have large amounts
to give.
First thing about AI is that it could radically reduce costs through
automation and thus make philanthropy advice a mass-market
service (possiblyeven the default).
Further advantages of using AI in this way are greater
personalisation and availability (i.e. 24/7/365).
Worth saying that many experts believe that predominant model
will be AI-augmented human advice in many fields,rather than full
automation, because people still like to have human interaction for
valuable transactions
HOWEVER:This may change in a future where AI is ubiquitous
and we are all at home with the idea of letting our lives be
governed by the advice and decisions of machines.
Making philanthropy advice more affordable and available still
leaves open question of what the nature of that advice is.
This is where I think things get really interesting.
OK, they don’t immediatelyget interesting.
First possibility is the pretty mundane one of using AI to give
advice on giving methods (i.e. vehicles available, tax info,
approaches etc.)
Basically the sort of info provided on many websites,and given by
existing advisers,but turned into AI-powered service.
3. Rhodri Davies- Programme Director, Giving Thought at Charities Aid Foundation
Next, let’s turn to more interesting applications
First, need to distinguish between2 approaches:
Maximising donor satisfaction
Maximising social outcomes/impact
OUTCOMES:
NEEDS: AI could be used to identify most pressing areas of
social/environmental need, based on analysis of huge data sets.
Data could come from existing public/private/vol sectorsets
In future, data could come from IoT/wearables
INTERVENTIONS:AI could be used to analyse data on social
impact, to find most effective ways of achieving given goals.
Both put huge onus on data:
Breaking down silos in public/private sector
Adopting open data approach in vol sector
Developing consistentsocial impact measurement
Harnessing potential of IoT
DONOR SATISFACTION
Subjective:just ask the donor what they want to achieve (old-
fashioned approach to phil advice)
BUT: self-reporting not always accurate (fast/slow thinking,
evidence about impact data reducing giving etc.)
Objective Social:Facebookalgorithm approach
Dangers: entrenching bias, strengthening filter bubbles,reinforcing
popular causes/organisations
4. Rhodri Davies- Programme Director, Giving Thought at Charities Aid Foundation
Objective Personal:What if in the future, AI could determine what
you should give to based on objective evidence of what gives you
most pleasure? (CF Harari, Homo Deus, Experiential/narrative
self)
What charities should do aboutthe above trends
Cross sectorforums (private sector,vol sec,govt)
Join debate over IA and put charity perspective (e.g. CAF Lords AI
committee response.)
AdoptOpenData approach (e.g. 360 Giving)
Develop social impact data
Work with IoT industry: both to harness data, and to develop M2M
giving