Artificial Intelligence.
Don’t believe the hype.
Super powerful computing and sophisticated maths models have led to AI wins at chess and the creation of tools like Chat GBT. Those who want to make money from it are making wild claims about its abilities. Most of it is not true. However, while AI isn’t going to take over the world there are genuine dangers in the way that it is designed and used for things it cannot do, and to amplify bias, discrimination and oppression.
“They should not be scared of the machines themselves, but of the people running the machines. People talk about creating robots that can think on their own, but the reality is far more mundane. And scarier.
Take the Universal Credit algorithm, which is creating terrible hardship for people even though it is simply doing what it was programmed to do. Automation has the capacity to scale up flawed policy decisions, causing incredible harm to thousands of people when not done properly.“
Stephanie Handcock, Human Rights Watch
The many weaknesses of AI
AI is powerful. You’ll have heard lots about this. But it is also fragile, makes huge amounts of errors, is full of bias and is energy and resource intensive. It does not function well in unstable environments and cannot understand humans.
” Much of what’s being sold as “AI” today is snake oil. It does not and cannot work.”
“Deployed AI systems often do not work. They can be constructed haphazardly, deployed indiscriminately, and promoted deceptively.”
“I think what IBM is excellent at is using their sales and marketing infrastructure to convince people who have asymmetrically less knowledge to pay for something”
AI bloopers: the ways in which AI go wrong help us understand what it is – and what it is not.
Experts on the latest round of AI Emperor’s New Clothes
The Fallacy of AI Functionality
AI is biased
Machine learning is sometimes sold as being more objective than humans. This ignores the reality that governments commission the programmes to do specific things, that programmers design the algorithms, that the training data uses the internet. Evidence shows that AI as it currently is actually amplifies and automates bias and inequality.
AI can’t predict children’s futures
Vast computing power does not mean that the impossible becomes possible. Robust evidence shows that AI cannot – of course – predict the future and that even huge amounts of data do not help.
“The study really highlights that at the end of the day, machine-learning tools are not magic”
Alice Xiang, Partnership on AI
“We can’t predict the future — that should be common sense. But we seem to have decided to suspend common sense when “AI” is involved.”
Conclusive proof that even with huge amounts of data the results are poor
It is mostly accepted that AI is poor at risk prediction but it has often been suggested that perhaps the poor results are because we don’t have enough data and therefore need more. An American study of the predictive ability of different risk prediction models blew this idea out of the water.
The ‘Fragile Families and child well-being study’ tracked the children of 4,000 families from birth, resulting in a huge amount of data on each child. Almost 13,000 data points.
Hundreds of researchers – with different programmes, algorithms and models – were asked to predict six life outcomes. None succeeded and the most complicated machine-learning techniques did no better than far simpler methods.
No model achieved better than missing 4 out of 5 children considered ‘at risk’ and wrongly flagging 3 out of 5 children not considered in need of help.
Gerd Gigerenzer
Using these systems for child protection is harming children and families
Children in need of help being missed while many who don’t are dragged through child protection processes
The shockingly bad results shown by the Fragile Families study are not just an academic exercise. Real families are being hurt by governments choosing to use algorithms even in the face of overwhelming evidence that they cannot be trusted.
AI bloopers:
Q. How do you confuse AI? A. Turn the picture on its side.
Dutch government brought down: 30,000 families wrongly accused of welfare fraud.
Societal harm
As well as the harm done to individual children and their families there are far reaching wider negative societal impacts.
- Increases inequality yet further
Puts existing bias and inequality on an automated – and therefore even larger – scale
- Loss of privacy
AI is data hungry and satisfying its needs means serious loss of privacy.
- Budgets used to police people rather than support
What more positive things could we do with all that money?!
- Transfer of power away from the people
Big data and AI gives the state more power over citizens but it also gives a lot of power to unaccountable companies.
- Loss of our protections
The government intends to remove many of our protections so as to smooth the path for AI
“The risks and harms have never been about “too powerful AI”. Instead: they’re about concentration of power in the hands of people, about reproducing systems of oppression…”
Professor Emily Bender
Despite flaws and failures the UK government hell bent on harmful AI and big data.
“Leveraging the whole public sector’s capacity to create demand for AI and markets for new services. The government has a clear role to play.“
UK Government
2022 National AI Strategy
Trojan horse for attacks on the poor
AI built on oppression
The world we need.
It doesn't have to be like this. We have a choice. We can choose rights for children, rebuilding our safety nets and supporting communities and what actually works for us, what makes us happy. We can reject the snake oil.
Michelle Bachelet
UN high commissioner, human rights
t often isn’t the either/or choice we are presented with – if we care about privacy we can have our cake and eat it – eg privacy concerns blocked covid tracking app. https://www.theguardian.com/world/2020/may/07/uk-coronavirus-contract-tracing-app-could-fall-foul-of-privacy-law-government-told
People – like Tony Blair – framed loss of liberties as a reasonable price for ‘beating’ covid https://www.bbc.co.uk/news/technology-52401763
But there were ways to do it without risking our privacy. But if you have a govt which is determined to get their hands on our data… Cartoon about contact tracing without loss of privacy https://ncase.me/contact-tracing/ https://www.technologyreview.com/2020/06/16/1002982/contact-tracing-without-big-brother/
.