Bartoletti's enlightening exploration of artificial intelligence demonstrates its immense power to positively and negatively impact society. An Artificial Revolution demonstrates the need for transparency, intervention, regulation and global governance to overhaul systems which are both pervasive and driven by human bias.
Ivana Bartoletti is a leading privacy and ethics consultant and visiting policy fellow at the Oxford Internet Institute at the University of Oxford. Her extensive work continues to shine a light on the need for equality, privacy and ethical consideration in the development and use of Artificial Intelligence. In An Artificial Revolution Bartoletti shows that many peoples fears about AI (sentient robots who overthrow the human race/my phone is listening to me etc.) are often misplaced or over-prioritised. More worrying is that not enough people are aware or concerned with the AI developments that are already being utilised and impacting societies around the world.
Importantly, Bartoletti is not advocating against AI or developments in technology; instead she is advocating for rigour, public consideration and global governance over something with phenomenal power to change the world. Her book highlights areas where this need for action is most urgent.
The Human Influence on AI
“We have internalised the idea that there is nothing more objective, neutral, informative and more efficient than data. This is misleading. When an algorithm is fed data a decision has already been made. Someone has already decided that some data should be chosen and other data should not. And if data is, in reality, people then some of us are being selected while others are being silenced.” - Ivana Bartoletti, An Artificial Revolution
A key aspect of Bartoletti's book which stood out to me was focusing on the human entities behind technology and how these can encode AI systems with inherent bias. As the quote above highlights, someone decides what is captured and therefore what is not. The data gathered can be used to influence or make decisions, and if AI is being used in areas like employment, healthcare, town planning, etc. , the consequences can be life altering. As a result, Bartoletti argues that AI systems can often be encoded with racial, classist and sexist bias, impacting enormous portions of society for the worse.
Some AI systems themselves can often include a huge amount of unseen human labour, much of which Bartoletti describes as 'exploitative and degrading' (Bartoletti, 2020:27). She gives the example of the factories of workers tasked with training machine learning systems to recognise indecent images. A machine doesn't know all the different ways child abuse can look unless humans undergo relentless hours of teaching it. Currently, the only way to do that is to view and label thousands of images as child abuse. The human workers who undergo this harrowing job do so for just a few pennies per label. The output we see is hailed as innovative tech with the growing ability to wipe child abuse imagery from social media, but behind the scenes are thousands of workers, underpaid and potentially traumatised by the process. The CEO takes home a few million, the workers take home a pittance and PTSD.
Surveillance and Decision Making
Most of us know that our device use collects data about us. Reading Bartoletti's book reminded me of a pop up shop that appeared at Old Street station one morning on my way to work a number of years ago. Set up by artist Ben Eine the Data Dollar store was a shop where you could trade in aspect of your data (photos, whatsapp messages etc.) in exchange for goods. Eine wanted to highlight that we are giving away our data free everyday and that others are profiting from it and, unlike in the Data Dollar Store, we receive nothing in return.
An Artificial Revolution shows some of the ways in which our data is potentially being used to surveil communities and potentially even profile and punish us. One example which stood out to me in the book was the discussions around using fitness data to try and manufacture correlations between individual's fitness tracker data and how likely they are to commit violent crime (Bartoletti, 2020:79). Searching for tenuous correlation between things like this seems ridiculous - yet it is ultimately profitable for someone to pursue it and sell this information on, whether it is based on scientific findings or not.
Politics and Opinion
And while our data is collected, we are consequently fed back a curated and personalised experience online. Our online presences silo us into likeminded groupings and create online ecosystems which reinforce our already held beliefs. An AI Revolution shows why this isn't always a good thing. Being fed a consistent diet of only one side of any story, experience, set of beliefs, political viewpoint or world view means you have to actively want to seek other voices and opinions. When Bartoletti puts this in the context of extremist beliefs for example, she captures why this silo effect can be dangerous.
Final Thoughts
An AI Revolution is a short yet informative book which leaves the reader with a lot of food for thought. Bartoletti's real world examples - both present and imminent - capture the issues with ungoverned AI systems which have significant impact on many facets of our society, from labour to domestic violence to political populism.
Importantly, Bartoletti calls for dialogue around technology to be 'nuanced'. Too often technological advancements have been presented in the binary of 'something that will solve all problems, or as something that will ultimately destroy our rights and freedoms' (Bartoletti, 2020:125) and the reality is that it is all in how it is utilised.
AI needs accountability, scrutiny and global norms wrapped around it.
An Artificial Revolution is recommended reading for those interested in dialogue around technology and are seeking level-headed input from an expert in the field. For anyone who has ever said 'I don't care what data they collect about me' it may also prove to be a worthwhile read!
Comentários