iHuman [Documentary]

Summary

The documentary examines the rapid growth of artificial intelligence (AI), its potential benefits, and the risks that arise when powerful systems are controlled by a small number of technology companies or governments. It argues that AI is not inherently good or bad; its effects depend on who controls it and what goals it serves.

How AI Works and Where It Is Heading

• Today’s AI is mostly narrow AI: systems designed for specific tasks, such as recommending online content, assessing insurance risk, reviewing job applications, or recognizing speech and images.
• Deep learning systems identify patterns in large datasets. The film highlights LSTM neural networks used for speech recognition, translation, and video recognition, including technologies embedded in smartphones.
• Researchers pursuing artificial general intelligence (AGI) aim to build systems that can perform a wide range of tasks and improve their own learning processes. The documentary raises concerns about the difficulty of predicting or controlling such systems, especially if they become more capable than humans.
• Potential benefits include medical advances, climate solutions, increased productivity, and automation. Potential costs include job displacement, inequality, and systems whose objectives do not align with human interests.

Data, Personalization, and Surveillance

• The film describes personal data as a major source of commercial and political power. Browsing history, location, purchases, social-media activity, language, and images can be analyzed to infer traits and predict behavior.
• Recommendation and advertising systems can personalize what people see, capture attention, and reinforce existing preferences. The documentary links these feedback loops to echo chambers, polarization, and the spread of political extremes.
• The film discusses the Cambridge Analytica controversy. Christopher Wylie describes how harvested Facebook data was used to profile personality traits and target voters. The documentary presents this as an example of data analysis being used to exploit psychological vulnerabilities.
• Facial-recognition systems can identify people and, according to a researcher interviewed, infer sensitive traits from faces. The film warns that these capabilities could threaten privacy and enable discrimination or persecution.

Uses in Policing, Government, and the Military

• Predictive policing uses historical crime data to estimate where crime may occur. The documentary warns that biased policing data can create a feedback loop: some communities receive more police attention, generating more recorded arrests and reinforcing the system’s original bias.
• Government and commercial databases can be combined to track people’s movements, relationships, and activities. The film describes this as enabling extensive surveillance, potentially including the reconstruction of a person’s past movements.
• The documentary examines AI’s military uses, including Project Maven, described as an effort to use machine learning to analyze drone video and identify potential battlefield targets. The project prompted objections and resignations among Google employees.
• Speakers warn that autonomous weapons and machine-speed warfare could reduce human oversight and increase the risk of lethal decisions being made by machines.

China, Social Scoring, and Competition

• The documentary describes China’s extensive use of mobile payments, facial recognition, and other data systems. It also discusses the government’s social-credit approach as a way of combining financial and political judgments about individuals.
• The film reports concerns about surveillance and predictive-policing systems in Xinjiang, including the use of technology to identify people for government scrutiny.
• AI development is framed as an international competition, particularly between the United States and China, with military and economic power among the stakes. The film questions whether regulation can keep pace with that race.

Deepfakes and Information Integrity

AI-generated or altered video can make people appear to say things they never said. The documentary notes potential creative uses but emphasizes the risks of manipulation, fake news, and the erosion of trust in online media.

Central Questions

The film repeatedly asks what kind of society people want AI to help create, how human rights and accountability can be protected, and whether international rules can limit surveillance and weaponization. It concludes that the technology’s future is not just a technical question but a political and ethical one.

Reviews, Guides, or Tutorials

None are presented. This is a documentary featuring interviews, commentary, and news footage.

Main Speakers and Sources

The documentary features AI researchers and developers, including an AGI researcher discussing neural networks and artificial curiosity, and an OpenAI-affiliated speaker discussing goal alignment. Other sources include academics studying data and facial analysis; technology and policy journalists; Cambridge Analytica whistleblower Christopher Wylie; Google employees and a former employee discussing Project Maven and Dragonfly; military and national-security commentators; and human-rights and surveillance researchers. Named figures include Peter Thiel and Eric Schmidt.