Human rights in the age of

HUMAN RIGHTS IN THE AGE OF ARTIFICIAL INTELLIGENCE
satellite imagery with weather and agronomic data to help farmers improve crop yields, diagnose and 
treat crop disease, and adapt to changing environments. This approach to farming is known as precision 
agriculture, and it can help increase farm productivity to feed more of the world’s growing population. 
Mitigating climate change, predicting natural disasters, and conserving wildlife: 
With the effects of 
climate change appearing around the world, machine learning is being used to make more accurate climate 
models for scientists. Already, AI is used to rank climate models and predict extreme weather events,
35
as 
well as to better predict extreme weather events and respond to natural disasters.
36
AI is also helpful for 
identifying and apprehending poachers and locating and capturing disease-spreading animals.
Making government services more efficient and accessible: 
Despite often being slow to adopt new 
technologies, governments around the world are using AI, from the local to the national levels, to make 
public services more efficient and accessible, with an emphasis on developing “smart cities.” AI is also 
being used to allocate government resources and optimize budgets.
37
HARMFUL AI
Perpetuating bias in criminal justice: 
There are many documented cases of AI gone wrong in the criminal 
justice system. The use of AI in this context often occurs in two different areas: risk scoring—evaluating 
whether or not a defendant is likely to reoffend in order to recommend sentencing and set bail—or so-called 
“predictive policing,” using insights from various data points to predict where or when crime will occur and 
direct law enforcement action accordingly.
38
In many cases, these efforts are likely well-intentioned. Use of 
machine learning for risk scoring of defendants is advertised as removing the known human bias of judges 
in their sentencing and bail decisions.
39
And predictive policing efforts seek to best allocate often-limited 
police resources to prevent crime, though there is always a high risk of mission creep.
40
However, the 
recommendations of these AI systems often further exacerbate the very bias they are trying to mitigate, either 
directly or by incorporating factors that are proxies for bias.