Start free trial
Trending New Genres Authors
EnglishEnglish
EspañolSpanish
简体中文Chinese
繁體中文Chinese (Traditional)
FrançaisFrench
DeutschGerman
日本語Japanese
PortuguêsPortuguese
ItalianoItalian
한국어Korean
РусскийRussian
NederlandsDutch
العربيةArabic
PolskiPolish
हिन्दीHindi
Tiếng ViệtVietnamese
SvenskaSwedish
ΕλληνικάGreek
TürkçeTurkish
ไทยThai
ČeštinaCzech
RomânăRomanian
MagyarHungarian
УкраїнськаUkrainian
IndonesiaIndonesian
DanskDanish
SuomiFinnish
БългарскиBulgarian
עבריתHebrew
NorskNorwegian
HrvatskiCroatian
CatalàCatalan
SlovenčinaSlovak
LietuviųLithuanian
SlovenščinaSlovenian
СрпскиSerbian
EestiEstonian
LatviešuLatvian
فارسیPersian
മലയാളംMalayalam
தமிழ்Tamil
اردوUrdu
Searching...
SoBrief
Code Dependent

Code Dependent

Living in the Shadow of AI
What AI hides: the workers, patients, and citizens whose lives reveal its real power.
by Madhumita Murgia 2024 309 pages
4.03
3k+ ratings
Listen
Immersive
V2.3
Amazon Kindle Audible
Try Full Access for 3 Days
Unlock listening & more!
Continue

Key Takeaways

Every AI system begins with human labor it prefers to hide

People label the world first. Machine learning is software that finds statistical patterns in large collections of examples. To learn what a pedestrian, traffic sign, or violent post looks like, systems depend on people to label and sort data. In Nairobi, Ian Koli can spend eight hours marking up an hour of road footage for driverless-car systems. Workers in Kenya, Bulgaria, and elsewhere also help prepare the data behind social media, shopping, and chatbots.

The human layer disappears. Workers may not know the client, the product, or how their task fits into the final system. Content moderators also help train software by judging material that platforms would rather users never see. The apparent autonomy of AI depends on a far-flung supply chain of human decisions.

Analysis

What's striking is that the labor is not merely physical support for a technical process. Annotators make interpretive choices about ambiguous images and instructions, so their judgment is part of the system's apparent intelligence. Hiding that contribution can hide responsibility as well as cost. The story also resists an easy conclusion: these jobs can provide income and digital skills, while still leaving workers with little visibility or influence. Counting jobs created is therefore an incomplete measure of whether an AI economy is fair.

Measure AI jobs by worker voice, not income alone

Opportunity can be real. Ian's work at Sama helped him move from precarious informal jobs into steady employment, connect his home to broadband, and attend college on a scholarship. Hiba, a refugee in Bulgaria, valued flexible annotation work that let her earn while caring for her family. Such jobs can bring income, skills, and room to imagine a different future.

Security has limits. When Hiba challenged a new shift rule, the company suspended her platform access for thirty days. The lost December income forced her family to borrow money. Separately, former content moderator Daniel Motaung alleged that Sama and Meta exploited workers and dismissed him for trying to organize; the companies disputed allegations in the ongoing case. His reported pay was about $2.20 per hour, and the work involved graphic content.

Analysis

Income and agency are separate dimensions of job quality. A job can improve a household's immediate choices while discouraging workers from questioning the rules that govern those choices. Labor economics often distinguishes having access to work from having bargaining power over its terms; these stories make that distinction concrete. The comparison should not erase the gains workers describe or assume every outsourcing job is abusive. It does suggest that evaluations of social-impact employers need to ask whether workers can raise concerns without risking their livelihood.

Use clinical AI as a second opinion, never as a substitute for care

Speed can save lives. At a remote Indian hospital, Dr Ashita Singh used an app to assess X-rays for tuberculosis. When a critically ill fourteen-year-old arrived, the app flagged possible TB and supported Ashita's decision to start treatment rather than wait days for a distant lab. The patient recovered. In a hospital without radiologists, a validated screening tool can shorten the time between an X-ray and a clinician's attention.

Context still matters. During the Covid-19 pandemic, the system sometimes confused tuberculosis with Covid because the lung images can look alike. Ashita reported the problem, and developers added a TB screening step. The case shows a productive relationship: clinicians can test a tool, identify its limits, and help improve it, while remaining responsible for diagnosis and treatment.

Analysis

This is a stronger model of medical automation than either replacing clinicians or dismissing software as inherently unsafe. The tool extends scarce diagnostic capacity, while a doctor supplies context, responsibility, and a relationship with the patient. The case also shows why deployment conditions matter: an accurate screen is useful only if someone can confirm the result and provide treatment. A system tested in one setting should not be presumed reliable elsewhere, especially when diseases, equipment, and patient populations differ.

Audit the model's proxy, not merely its accuracy

A neat measure can mislead. A health algorithm used by a major US provider estimated patients' needs using their annual healthcare costs. That seemed reasonable because illness often leads to more spending. But Black patients may receive less care because of access barriers and discrimination, so lower costs did not mean lower need. Researchers found the system substantially under-prioritized Black patients for additional support. Replacing cost with a better measure of health needs reduced the bias.

Proxies carry history. A proxy is a stand-in measure used when the desired quality is difficult to observe directly. The model did not need a race variable to produce racial inequality: it learned from a measure shaped by unequal access to care. Similar risks arise whenever a system treats recorded activity as a neutral picture of people's circumstances.

Analysis

The key audit question is not only whether a model predicts its chosen target accurately, but whether that target deserves to guide the decision. Accuracy can be mathematically impeccable and socially misdirected. This distinction is useful well beyond healthcare: spending, arrests, or past performance may reflect unequal opportunity as much as underlying need or ability. Yet proxy choices cannot be settled by technical testing alone. They involve judgments about which errors matter most and who bears the cost of being misclassified.

Predictive systems can turn yesterday's policing into tomorrow's suspicion

Arrests became a forecast. Amsterdam's ProKid system used police records and other information to identify children considered at risk of future crime. Its training data included arrests, not proof that a person had committed an offense. Because police attention was already concentrated on some immigrant communities, more contact could create more records, which could then reinforce the system's suspicion.

A label changes behavior. Diana Sardjoe's sons were placed on city lists, including one informed by ProKid. Her family described intrusive visits, stigma, and difficulty challenging the decision. Her older son, who had not been convicted of a crime, began offending after the label and surveillance damaged his sense of possibility, according to Diana. A reported decline in arrests across the program did not establish that the algorithm caused improvement or that the harms were justified.

Analysis

Prediction is not passive when institutions act on the score. A risk label can alter the opportunities and scrutiny a person receives, so the resulting data may partly record the effect of the intervention itself. This is a feedback problem, not just a technical flaw. Early help can be valuable, but support becomes coercive when families cannot understand why they were selected, refuse intervention, or correct the record. The book's emphasis on forgiveness adds an important question: should a child's past remain a permanent forecast?

A fabricated intimate image can inflict a real, lasting injury

The body can be falsified. A deepfake is a realistic image or video generated or manipulated with AI. Poet Helen Mort discovered images that used her face in fabricated scenes of sexual violence. Australian lawyer Noelle Martin found doctored sexual images of herself online and later discovered AI-generated videos. Neither had consented to their creation or distribution, and removing one copy could not stop others from spreading.

Scale outruns recourse. A 2019 study cited in the book estimated that roughly 95 percent of online deepfake videos were non-consensual pornography, with women making up nearly all targets. Laws and platform responses were uneven, leaving victims to seek takedowns or pursue difficult legal routes. Noelle's advocacy helped change Australian law to cover altered intimate images, showing that public testimony can turn private harm into a policy issue.

Analysis

The important distinction is between whether an image is authentic and whether the violation is real. Fabrication does not make the loss of privacy, safety, or reputation imaginary. Digital copies also change the time horizon of abuse: an image can reappear after the original incident has faded from public attention. Legal remedies need to address creation, distribution, and persistence, while protecting ordinary expression. A purely technical response, such as better detection, cannot decide who should bear responsibility or provide support to victims.

Face recognition can make public space a checkpoint for ordinary life

Identification changes the street. Facial-recognition systems compare a face with stored images, but errors are uneven. The Gender Shades study found error rates below 1 percent for light-skinned men and as high as 35 percent for darker-skinned women in systems it tested. Misidentifications have contributed to wrongful arrests. Even a correct match can enable officials to track people who are gathering, protesting, or simply moving through public space.

Surveillance changes behavior. In London, artists formed the Dazzle Club and used face paint designed to confuse recognition software, turning surveillance into a visible public debate. In other settings described in the book, camera networks were used to identify protesters and monitor marginalized communities. The issue is not only whether a system can recognize a face, but who controls the records and what consequences follow.

Analysis

Accuracy and legitimacy are different tests. A perfectly accurate system could still make public life less free if people are identified without notice, consent, or a meaningful reason. Conversely, a system with unequal error rates can expose some groups to greater risk of detention or suspicion. The Dazzle Club's action makes an abstract question embodied: people experience surveillance as a change in how they can occupy shared space. Public safety claims therefore need limits, independent scrutiny, and clear rules about retention and use.

Give algorithm-managed workers the data and leverage to challenge their boss

The app sets the terms. Delivery and ride-hailing platforms use algorithms to assign work, calculate pay, assess performance, and flag possible fraud. Courier Armin Samii built UberCheats after suspecting the app had undercounted his delivery distances. Across about 6,000 trips entered by users, 17 percent appeared underpaid, by an average of 1.35 miles per trip. Drivers could not readily check the company's calculations themselves.

Workers can compare notes. Couriers share tips, organize protests, and pool earnings data through groups and worker-led projects. In the UK, a Supreme Court ruling recognized Uber drivers as workers entitled to protections such as minimum wage and holiday pay. These examples show that opacity is not inevitable: access to records, a human appeal, and collective organization can give workers ways to contest automated decisions.

Analysis

Algorithmic management turns an employment relationship into a dispute over information. The platform sees patterns across thousands of trips; an individual driver sees a payment and a vague explanation. Collective data gathering helps close that asymmetry, while labor law can clarify that software does not erase an employer's obligations. Not every pay discrepancy proves deliberate underpayment, and the tool did not explain every error. Its value was making a hidden calculation testable and giving workers evidence to discuss together.

Treat fluent AI as a fallible pattern engine, not a knowing expert

Fluency can imitate certainty. Large language models generate text by predicting likely word sequences from patterns in training data. They are not search engines that verify facts, and they can invent names, cases, or sources. New York lawyer Steven Schwartz cited fictional cases generated by ChatGPT because he mistook the system for a research tool. A confident answer is not evidence that the system understands or checked its claims.

The inputs have consequences. Generative systems are trained on vast collections of human writing, art, images, and voices. The book describes creative workers losing assignments or being asked to revise cheaper AI-generated work built from material like theirs. A chemistry red-team exercise also showed that a model could lower the practical barriers to designing a dangerous compound. Useful output and serious risk can come from the same pattern-making capacity.

Analysis

The word understanding invites users to attribute intention and reliability where neither is guaranteed. Treating language models as prediction systems makes their strengths clearer: rapid drafting, recombination, and assistance with routine tasks. It also clarifies where verification is essential, especially in law, medicine, and safety-sensitive work. The creative labor issue raises a separate question from accuracy: even a technically impressive output can depend on contested use of other people's work. Better tools do not settle questions of consent, compensation, or authorship.

Build a route to appeal before automating consequential decisions

People need a way back in. Lawyer Cori Crider's group, Foxglove, challenged UK systems used to sort visa applications and assign exam grades. In 2020, protests and legal pressure helped overturn an exam-grading algorithm that disadvantaged many students from lower-performing schools. Workers have also taken platforms to court over automated pay deductions and suspensions, winning some access to personal data and human review.

Rights require practical handles. A person affected by an automated decision needs to know that automation was involved, understand the basis for the decision, correct relevant information, and reach someone empowered to reconsider it. A disclaimer or nominal human reviewer is not enough if that person cannot change the outcome. Legal action, unions, journalism, and community groups can turn isolated complaints into public accountability.

Analysis

Appeal is not an administrative afterthought. It is a safeguard against the inevitable mismatch between a statistical category and an individual life. Due process principles such as notice, explanation, and a chance to contest a decision become especially important when automated systems are difficult to inspect. The cases also show the limits of litigation: legal victories can open a door without revealing every model rule or changing the wider business incentives. Effective accountability needs both individual remedies and collective oversight.

Judge AI by whether it protects human dignity, not by novelty

Values need public authors. In Rome, religious leaders, technology companies, and public institutions joined the Rome Call, a voluntary pledge proposing ethical principles for AI. These included explainability, inclusion, privacy, and accountability. The book argues that a company-written set of rules cannot represent humanity by itself, particularly when people affected by AI have little influence over its design.

Ask before deploying. The practical test is not simply whether a system works, but who benefits, who bears its errors, and whether affected people can choose a human decision-maker or challenge the result. Could a health tool reach the patients whose data helped build it? Can workers see how pay is calculated? Can a person remove intimate images or contest a risk label? These questions put dignity and agency ahead of technological spectacle.

Analysis

The book's most durable standard is not pro-AI or anti-AI. It is whether a system expands people's real capabilities without making them easier to exploit, classify, or silence. That standard accommodates Ashita's diagnostic tool and condemns opaque systems that punish people without recourse. Voluntary ethical pledges can create shared language, but they cannot substitute for enforceable rules or public participation. The difficult work is deciding whose values count, how conflicts are resolved, and who remains responsible when an automated decision causes harm.

Analysis

Code Dependent is a reported investigation built from lives encountered at the edge of technological power: data annotators in Nairobi, a mother in Amsterdam, a rural doctor in India, delivery workers, artists, and people subjected to surveillance. Its central argument is not that AI is uniformly harmful. The same pattern-finding capacity that can help a clinician identify tuberculosis can reinforce racial bias in healthcare, or turn a police record into a forecast of future criminality. The decisive variable is how the system is built, where it is used, and who can contest its consequences.

The book's strongest conceptual move is to treat AI as an institution rather than an isolated invention. Models depend on labor, data, public infrastructure, legal rules, and organizational incentives. Their apparent independence can obscure those dependencies, making power harder to locate. This also explains why technical accuracy is insufficient as a measure of social value. A model may predict a narrow target well while the target itself encodes unequal treatment, or while the intervention creates harms that the metric does not count.

The contrast between Ashita's clinical tool and Amsterdam's youth risk scoring is especially instructive. Ashita retains expertise, context, and responsibility, and the software supports rather than displaces judgment. The families targeted by predictive policing had far less information or ability to refuse. The difference is not simply human versus machine. It is whether an affected person has a meaningful relationship to the decision process.

The account does not fully resolve how to balance beneficial automation against the danger of concentrated control. But its reporting makes that tension concrete rather than abstract. Its answer is institutional: disclose, test, compensate, provide appeal, and involve affected communities before systems become infrastructure. Human dignity is not a decorative ethical value added after deployment. It is a test of whether the system deserves authority at all.

Summary in 30 Seconds
AI depends on hidden human labor, including annotation work sometimes paid near $2.20 an hour. Accuracy can conceal bias when systems use unequal healthcare spending or arrest records as proxies for need or danger. Facial recognition errors reached 35% for darker-skinned women in one study. About 95% of deepfake videos were estimated to be nonconsensual pornography. Effective safeguards require transparency, worker power, human review, and meaningful appeals.
Contains spoilers

Last updated:

Report Issue

Review Summary

4.03 out of 5
Average of 3k+ ratings from Goodreads and Amazon.

Code Dependent by Madhumita Murgia explores the impact of AI on society, focusing on marginalized communities and ethical concerns. Readers praise its accessibility, compelling storytelling, and balanced perspective on AI's potential benefits and risks. The book highlights issues like data colonialism, algorithmic bias, and the exploitation of workers in the AI industry. While some critics note a lack of technical depth and solutions, most commend Murgia's ability to humanize complex topics and provoke critical thinking about our AI-driven future.

Your rating:
4.52
185 ratings
Want to read the full book?

Glossary

data colonialism

Extracting value from lives

A framework for understanding how powerful institutions convert people's everyday activities, relationships, and personal information into resources for economic or political use. The book connects this extraction to older colonial patterns: value and control accrue to distant organizations, while affected communities often have little say over the data, the systems built from it, or who benefits.

ghost work

Hidden human AI labor

Human labor that supports digital services while remaining largely invisible to users and sometimes to the workers themselves. In AI, this includes labeling images, sorting text, and moderating content so systems can be trained or operated. The term emphasizes that apparently automated products often depend on dispersed workers whose contribution and working conditions are concealed.

automation bias

Overreliance on automated advice

A tendency for people to trust automated outputs too readily or to defer to them when making decisions. In the book's medical example, a clinician could become less attentive to patient history or examination if an AI score is treated as authoritative. The risk is not that automation is always wrong, but that users may stop checking its assumptions and limits.

deepfake

AI-generated or altered media

A realistic image, audio recording, or video made or manipulated using artificial intelligence. The book focuses especially on non-consensual sexual imagery, where someone's face or likeness is placed into fabricated intimate scenes. Because such material can be produced and copied easily, victims may face lasting harm even when a particular website removes one version.

ProKid

Dutch youth risk assessment

A machine-learning system developed with Dutch police to estimate whether young people might be arrested in the future. It used police records and other information, including contacts involving a child, family members, or peers. The book stresses that arrest data is not the same as proof of criminal behavior, and that intervention based on a score can intensify surveillance and stigma.

Integrated Joint Operations Platform (IJOP)

Xinjiang surveillance system

A data platform used by authorities in China's Xinjiang region to combine information about people's movements, devices, biometrics, relationships, and behavior. Human Rights Watch researcher Maya Wang examined a related app and found indicators that could flag ordinary religious or social practices as suspicious. The book describes the platform as part of surveillance and detention targeting Uyghur Muslims.

large language model (LLM)

Text pattern prediction system

A type of AI trained on large collections of text to generate likely word sequences in response to prompts. LLMs can produce fluent explanations, stories, or code, but fluency does not ensure factual accuracy or human understanding. They may fabricate details, sources, or cases, so consequential claims need independent verification.

algor-ethics

Ethics built into algorithms

A term used in the Rome Call to describe the effort to embed human values and ethical safeguards into AI design and deployment. The proposal emphasizes that systems used in decision-making should respect human dignity and call for human involvement when values or safety are uncertain. It is presented as a shared ethical framework, not an enforceable law.

Rome Call

Voluntary AI ethics pledge

A pledge initiated by Vatican adviser Paolo Benanti to bring technology companies, public institutions, and religious representatives into discussion about responsible AI. Its principles include explainability, inclusion, privacy, and accountability. The book presents it as an attempt to widen the voices shaping AI values, while noting that voluntary signatures do not legally compel organizations to comply.

Gender Shades

Facial analysis bias study

A study by Joy Buolamwini, Timnit Gebru, and collaborators that compared the accuracy of commercial facial-analysis systems across skin tone and gender groups. The results reported substantially higher error rates for darker-skinned women than for light-skinned men in the systems tested. The book uses the study to show that aggregate accuracy can conceal unequal risks for particular populations.

FAQ

What's Code Dependent: Living in the Shadow of AI about?

  • Exploration of AI's Impact: The book examines how AI is reshaping various aspects of human life, including work, health, identity, and freedom. It delves into the consequences of our increasing reliance on AI technologies.
  • Personal Stories: Author Madhumita Murgia shares real-world encounters with AI, highlighting the experiences of individuals affected by these technologies. This narrative approach makes the implications of AI more tangible and relatable.
  • Critical Perspective: Murgia critiques the power dynamics and ethical concerns surrounding AI, particularly how it can perpetuate inequalities and affect marginalized communities. The book serves as a call to action for greater awareness and accountability in AI development.

Why should I read Code Dependent: Living in the Shadow of AI?

  • Timely Topic: As AI becomes more integrated into daily life, understanding its implications is crucial. This book provides insights into how AI affects personal and societal levels.
  • Engaging Narratives: Murgia's use of personal stories makes complex topics accessible and engaging. Readers can connect with the human experiences behind the technology.
  • Critical Analysis: The book encourages readers to think critically about the ethical and social implications of AI. It challenges the notion of technology as inherently beneficial, prompting reflection on its consequences.

What are the key takeaways of Code Dependent: Living in the Shadow of AI?

  • AI's Ubiquity: AI is now embedded in many aspects of life, from healthcare to employment, influencing decisions that affect our daily existence. Understanding this integration is essential for navigating modern society.
  • Data Colonialism: The book discusses the concept of data colonialism, where marginalized communities' data is exploited for profit without their consent or benefit. This highlights the need for ethical considerations in data usage.
  • Human Agency: Murgia emphasizes the importance of maintaining human agency in the face of automated systems. The book advocates for a balance between technological advancement and ethical responsibility.

What are the best quotes from Code Dependent: Living in the Shadow of AI and what do they mean?

  • "Ultimately, we are the products.": This quote underscores the idea that our data is commodified by tech companies, turning individuals into products for profit. It highlights the need for awareness about how personal information is used.
  • "Our blindness to how AI systems work means we can’t properly comprehend when they go wrong.": This statement reflects the dangers of relying on opaque AI systems without understanding their mechanisms. It calls for transparency and accountability in AI development.
  • "The power is concentrated in the hands of a few companies.": This quote points to the monopolistic tendencies in the tech industry, where a small number of corporations control vast amounts of data and influence. It raises concerns about equity and access in the digital age.

How does Code Dependent: Living in the Shadow of AI address the concept of data colonialism?

  • Definition of Data Colonialism: Murgia describes data colonialism as the exploitation of marginalized communities' data for profit, often without their knowledge or consent. This concept critiques the power dynamics inherent in data collection.
  • Impact on Communities: The book illustrates how data colonialism perpetuates existing inequalities, particularly affecting vulnerable populations. It emphasizes the need for ethical data practices that respect individuals' rights.
  • Call for Action: Murgia advocates for greater awareness and regulation of data practices to protect marginalized communities. The book encourages readers to consider the implications of their data being used in AI systems.

How does Code Dependent: Living in the Shadow of AI explore the relationship between AI and healthcare?

  • AI in Medical Diagnostics: Murgia discusses how AI technologies are being used to improve medical diagnostics and patient care. This includes applications in identifying diseases and personalizing treatment plans.
  • Ethical Concerns: The book raises questions about the ethical implications of AI in healthcare, particularly regarding data privacy and the potential for bias in algorithms. It emphasizes the need for equitable access to AI-driven healthcare solutions.
  • Human Oversight: Murgia argues for the importance of maintaining human oversight in medical decisions, even as AI becomes more prevalent. The book advocates for a collaborative approach between technology and healthcare professionals.

What are the implications of AI on personal identity as discussed in Code Dependent: Living in the Shadow of AI?

  • Identity and Surveillance: Murgia explores how AI technologies, particularly facial recognition, can impact personal identity and privacy. The book discusses the risks of being constantly monitored and the effects on individual autonomy.
  • Data Representation: The book highlights how individuals are often reduced to data points in AI systems, which can misrepresent their identities. This raises concerns about the accuracy and fairness of algorithmic decisions.
  • Cultural Impact: Murgia examines how AI can reinforce stereotypes and biases, affecting how individuals are perceived based on their identity. The book calls for a more nuanced understanding of identity in the context of AI.

How does Code Dependent: Living in the Shadow of AI address the future of AI?

  • Predictions and Trends: Murgia discusses emerging trends in AI, including generative AI and its potential to reshape creative industries. The book speculates on how these advancements may further alter human experiences.
  • Societal Changes: The book explores the broader societal implications of AI, including shifts in employment, education, and social interactions. Murgia emphasizes the need for proactive measures to address these changes.
  • Ethical Frameworks: Murgia advocates for the development of ethical frameworks to guide the future of AI. The book encourages readers to consider how society can shape technology to serve humanity rather than the other way around.

What are the challenges of AI implementation highlighted in Code Dependent: Living in the Shadow of AI?

  • Technical Limitations: Murgia discusses the technical challenges of implementing AI systems, including issues of accuracy and reliability. These limitations can hinder the effectiveness of AI in various applications.
  • Bias and Inequality: The book highlights the risk of bias in AI algorithms, which can perpetuate existing inequalities. Murgia calls for greater scrutiny of data sources and algorithmic design to mitigate these risks.
  • Public Trust: Murgia emphasizes the importance of building public trust in AI technologies. The book argues that transparency and accountability are essential for fostering confidence in AI systems.

How does Code Dependent: Living in the Shadow of AI suggest we reclaim agency in an AI-driven world?

  • Awareness and Education: Murgia advocates for increased awareness and education about AI technologies and their implications. Understanding how AI works is crucial for individuals to navigate its effects on their lives.
  • Collective Action: The book encourages collective action to demand ethical practices in AI development and implementation. Murgia believes that communities can work together to hold companies accountable.
  • Empowerment through Knowledge: Murgia emphasizes the importance of empowering individuals with knowledge about their data and rights. This empowerment can help people reclaim agency in an increasingly automated world.

How does Code Dependent: Living in the Shadow of AI address the issue of AI in the workplace?

  • Gig Economy Focus: The book explores the experiences of gig workers who navigate precarious employment conditions exacerbated by AI systems. Murgia highlights how algorithms dictate job allocation and pay, often to the detriment of workers.
  • Human Stories: Through personal narratives, the author illustrates the emotional and financial toll of working in an AI-driven gig economy. These stories provide a human face to the statistics and trends surrounding gig work.
  • Advocacy for Fair Practices: Murgia calls for fair wages and better working conditions for gig workers, emphasizing the need for collective action and solidarity among workers. She argues that ethical considerations must be integrated into the design of AI systems that impact labor.

What role do activists play in the narrative of Code Dependent: Living in the Shadow of AI?

  • Voices of Resistance: The book features activists who challenge the status quo and advocate for ethical AI practices. Their stories highlight the importance of grassroots movements in pushing for change.
  • Collective Action: Murgia emphasizes the power of collective action in addressing the challenges posed by AI and data colonialism. Activists work to raise awareness and mobilize communities to demand accountability from tech companies.
  • Empowerment through Advocacy: The author illustrates how activism can empower marginalized communities to reclaim their rights and advocate for their needs. This includes fighting for fair treatment in the gig economy and demanding transparency in AI systems.

About the Author

Madhumita Murgia is an accomplished journalist and expert on artificial intelligence. As the AI Editor for the Financial Times, she has established herself as a leading voice in technology reporting. Madhumita Murgia is known for her ability to break down complex AI concepts for a general audience, combining technical knowledge with a focus on the human impact of technology. Her work frequently appears in publications like Wired, where she has covered topics such as data brokers and emerging tech trends. Murgia's background in investigative journalism and her global perspective inform her nuanced approach to exploring the societal implications of AI.

Download PDF

To save this Code Dependent summary for later, download the free PDF. You can print it out, or read offline at your convenience.
Download PDF
File size: 0.44 MB     Pages: 23

Download EPUB

To read this Code Dependent summary on your e-reader device or app, download the free EPUB. The .epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.
Download EPUB
File size: 1.45 MB     Pages: 30
Want to read the full book?
Follow
Listen
Home
Swipe
Library
Get App
Try Full Access for 3 Days
Listen, bookmark, and more
Compare Features Free Pro
📖 Read Summaries
Read unlimited summaries. Free users get 3 per month
🎧 Listen to Summaries
Listen to unlimited summaries in 40 languages
—
❤️ Unlimited Bookmarks
Free users are limited to 4
—
📜 Unlimited History
Free users are limited to 4
—
📥 Unlimited Downloads
Free users are limited to 1
—
Risk-Free Timeline
Today: Get Instant Access
Listen to full summaries of 26,000+ books. That's 12,000+ hours of audio!
Day 2: Trial Reminder
We'll send you a notification that your trial is ending soon.
Day 3: Your subscription begins
You'll be charged on Nov 24,
cancel anytime before.
Consume 2.8× More Books
2.8× more books Listening Reading
Our users love us
1M+ readers
Trustpilot Rating
TrustPilot
4.6 Excellent
This site is a total game-changer. I've been flying through book summaries like never before. Highly, highly recommend.
— Dave G
Worth my money and time, and really well made. I've never seen this quality of summaries on other websites. Very helpful!
— Em
Highly recommended!! Fantastic service. Perfect for those that want a little more than a teaser but not all the intricate details of a full audio book.
— Greg M
Save 62%
Yearly
$119.88 $44.99/year/yr
$3.75/mo
Monthly
$9.99/mo
Start a 3-Day Free Trial
3 days free, then $44.99/year. Cancel anytime.
Unlock a world of fiction & nonfiction books
26,000+ books for the price of 2 books
Read any book in 10 minutes
Discover new books like Tinder
Request any book if it's not summarized
Read more books than anyone you know
#1 app for book lovers
Lifelike & immersive summaries
30-day money-back guarantee
Download summaries in EPUBs or PDFs
Cancel anytime in a few clicks

Why one summary?

Scanner
Find a barcode to scan

⭐
⭐
⭐
We have a special gift for you
Open
38% off
DISCOUNT FOR YOU
⭐
⭐
⭐
$79.99
$49.99/year
Only $4.16 per month
Continue
2 taps to start, super easy to cancel
Settings
General
Widget
Loading...