Review: “Code Dependent: Living in the Shadow of AI” by Madhumita Murgia (audiobook edition)
This is a good book looking at the social impacts of artificial intelligence by telling international stories – something that has been missing so far in my reading about AI. It should be the starting point for anyone looking at the wider repercussions of this technology which has taken large strides forward in the last couple of years.
It starts by looking at people in poorer countries who are being used by the tech giants to manually classify items within images so that they can be used to train large language models. This process is not artificial. Whilst it provides employment in the short term, it will ultimately mean that models will improve their abilities and remove the need for these jobs in the future.
Healthcare is one of the areas where AI is expected to make the biggest impact. It is already aiding medical professionals to diagnose illnesses and helping people in areas where there are few medical resources. Care will have to be taken to make sure that assumptions by the creators of these systems do not result in oversights which would impact their ability to help women, people of colour and the economically poor: those who are traditionally the most overlooked. AI can definitely play an important role in healthcare but it should be in addition to the human factor – the people on the ground dealing with the population at large, not just doctors and nurses but also the other less formal roles which can contribute considerably to better health outcomes.
An interesting point that I hadn’t come across before is that extremely personal information is being gathered from people and this is being used by global corporations. In the past this information may have only been available to governments. Often people are not aware what the data will be used for and the final products may not be available to them. The aims may be laudable – reducing teenage pregnancy, raising women out of poverty – but rather than some new computer system what they could really do with is better childcare.
The scope of the book is very broad. I have focused on healthcare but it considers deepfakes and the use of tools to deliberately cause harm, the use of facial recognition software as a surveillance tool for police, governments and private companies, the use of AI for repression of minorities or oversight of an entire population, the rise of the gig economy and essentially working for an app, the impact of being identified as someone likely to commit crime in the future, generative AI hallucinating and the threat to people’s livelihoods.
The author does a very solid reading of her book – given that these are personal stories she gets across the humanity and her personality comes over as warm, interested and engaged. I enjoy women narrators and there aren’t enough of them. A good recording too – no changes in sound, volume or retouches.
Spare Cycles 