Manifesto on algorithmic humanitarianism

Dan McQuillan

The nature of machine learning operations mean they will actually deepen some humanitarian problematics and introduce new ones of their own. This banality of machine learning is also its power.

intro

Artificial Intelligence (AI) is undergoing a period of massive expansion.

This is not because computers have achieved human-like consciousnesss but because of advances in machine learning, where computers learn from training data how to classify new data.

At the cutting edge are the neural networks that have learned to recognise human faces or play Go.

Recognising patterns in data can be used as a predictive tool, and AI is being applied to echocardiograms to predict heart disease

to workplace data to predict if employees are going to leave

and to social media feeds to detect signs of incipient depression or suicidal tendencies.

One activity that seems distant from AI is humanitarianism;

the organisation of on-the-ground aid to fellow human beings in crisis due to war, famine or other disaster.

But humanitarian organisations, too, will adopt AI because it seems able to answer questions at the heart of humanitarianism

such as 'who should we save?' and 'how can we be effective at scale?'

It resonates strongly with existing modes of humanitarian thinking and doing

in particular the principles of neutrality and universality.

The way machine learning consumes big data and produces predictions

suggests it can both grasp the enormity of the humanitarian challenge and provide a data-driven response.

But the nature of machine learning operations mean they will actually deepen some humanitarian problematics

and introduce new ones of their own.

Thinking about how to avoid this raises wider questions about emancipatory technics

and what else needs to be in place to produce machine learning for the people.

maths

There is no intelligence in Artificial Intelligence

nor does it really learn, even though it's technical name is machine learning

it is simply mathematical minimisation

Like at school, fitting a straight line to a set of points

you pick the line that minimises the differences overall

Machine learning does the same for complex patterns

it fits input features to known outcomes by minimising a cost function

the fit is a model that can be applied to new data to predict the outcome

The most influential class of machine learning algorithms are neural networks

which is what startups call 'deep learning'

They use backpropagation: a minimisation algorithm that produces weights in different layers of neurons

anything that can be reduced to numbers and tagged with an outcome can be used to train a model

the equations don't know or care if the numbers represent amazon sales or earthquake victims

This banality of machine learning is also its power

it's a generalised numerical compression of questions that matter

there is no comprehension within the computation

the patterns are correlation not causation

the only intelligence comes in the same sense as military intelligence; that is, targeting

But models produced by machine learning can be hard to reverse into human reasoning

why did it pick this person as a bad parole risk? what does that pattern of weights in the 3rd layer represent? we can't necessarily say.

reasoning

Machine learning doesn't just make decisions without giving reasons, it modifies our very idea of reason

that is, it changes what is knowable and what is understood as real

It operationalises the two-world metaphysics of neoplatonism

that behind the world of the sensible is the world of the form or the idea.

A belief in a hidden layer of reality which is ontologically superior,

expressed mathematically and apprehended by going against direct experience.

Machine learning is not just a method but a machinic philosophy

What might this mean for the future field of humanitarian ai?

It makes machine learning prone to what Miranda Fricker calls epistemic injustice

She meant the social prejudice that undermines a speaker's word

but in this case it's the calculations of data science that can end up counting more than testimony

The production of opaque predictions with calculative authority

will deepen the self-referential nature of the humanitarian field

while providing a gloss of grounded and testable interventions

Testing against unused data will produce hard numbers for accuracy and error

while making the reasoning behind them inaccessible to debate or questioning

Using neural networks will align with the output driven focus of the logframe

while deepening the disconnect between outputs and wider values

Hannah Arendt said many years ago that cycles of social reproduction have the character of automatism.

The general threat of AI, in humanitarianism and elsewhere, is not the substitution of humans by machines but the computational extension of existing social automatism

production

Of course the humanitarian field is not naive about the perils of datafication

We all know machine learning could propagate discrimination because it learns from social data

Humanitarian institutions will be more careful than most to ensure all possible safeguards against biased training data

but the deeper effect of machine learning is to produce new subjects and to act on them

Machine learning is performative, in the sense that reiterative statements produce the phenomena they regulate

Humanitarian AI will optimise the impact of limited resources applied to nearly limitless need

by constructing populations that fit the needs of humanitarian organisations

This is machine learning as biopower

it's predictive power will hold out the promise of saving lives

producing a shift to preemption

but this is effect without cause

The foreclosure of futures on the basis of correlation rather than causation

it constructs risk in the same way that twitter determines trending topics

the result will be algorithmic states of exception

According to Agamben, the signature of a state of exception is ‘force-of’

actions that have the force of law even when not of the law

Logistic regression and neural networks generate mathematical boundaries

but cybernetic exclusions will have effective force by allocating and withholding resources

a process that can't be humanised by having a humanitarian-in-the-loop

because it is already a technics, a co-constituting of the human and the technical

decolonial

The capture, model and preempt cycle of machine learning will amplify the colonial aspects of humanitarianism

unless we can develop a decolonial approach to its assertions of objectivity, neutrality and universality

We can look to standpoint theory, a feminist and post-colonial approach to science

which suggests that positions of social and political disadvantage can become sites of analytical advantage

this is where our thinking about machine learning & AI should start from

but I don't mean by soliciting feedback from humanitarian beneficiaries

Participation and feedback is already a form of socialising subjects

and with algorithmic humanitarianism every client interaction will be subsumed into training data

They used to say 'if the product is free, you are the product'

but now, if the product is free, you are the training data

training for humanitarian AI and for the wider cybernetic governance of resilient populations

Machine learning can break out of this spiral through situated knowledge

as proposed by Donna Daraway as a counterweight to the scientific ‘view from nowhere’,

a situated approach that is not optional in its commitment to a particular context

How does machine learning look from the standpoint of Haiti's post-earthquake rubble or from an IDP camp

No refugee in a freezing factory near the Serbian border with Croatia is going to be signing up for Andrew Ng's MOOC on machine learning any time soon

How can democratic technics be grounded in the humanitarian context?

people's councils

It may seem obvious that if machine learning can optimise Ocado deliveries then it can help with humanitarian aid

but the politics of machine learning are processes operating at the level of the pre-social

One way to counter this is through popular assemblies and people's councils

bottom-up, confederated structures that implement direct democracy

replacing the absence of a subject in the algorithms with face-to-face presence

contesting the opacity of parallel computation with open argument

and the environmentality of algorithms with direct action

The role of people's councils is not to debate for its own sake

but the creation of alternative structures, in the spirit of gustav landauer's structural renewal

An emancipatory technics is one that co-constitutes active agents and their infrastructures

As Landauer said, people must 'grow into a framework, a sense of belonging, a body with countless organs and sections'

as evidenced in Calais, where people collectively organised warehouse space, van deliveries and cauldrons to cook for 100s, while regularly tasting tear gas

I suggest that solidarity is an ontological category prior to subject formation

collective activity is the line of flight from a technological capture that extends market relations to our intentions

It is a politics of becoming – a means without end to counter AI's effect without cause

close

In conclusion

as things stand, machine learning and so-called AI will not be any kind of salvation for humanitarianism

but will deepen the neocolonial and neoliberal dynamics of humanitarian institutions

But no apparatus is a closed teleological system; the impact of machine learning is contingent and can be changed

it's not a question people versus machines but of a humanitarian technics of mutual aid

In my opinion this requires a rupture with current instantiations of machine learning

a break with the established order of things of the kind that Badiou refers to as an event

the unpredictable point of excess that makes a new truth discernible

and constitutes the subjects that can pursue that new truth procedure

The prerequisites will be to have a standpoint, to be situated, and to be committed

it will be as different to the operations of Google as the Balkan aid convoys of the 1990s were to the work of the ICRC

On the other hand, if an alternative technics is not mobilised,

the next generation of humanitarian scandals will be driven by AI.

Presented at the symposium on 'Reimagining Digital Humanitarianism', Goldsmiths, University of London, Feb 16th 2018./OD