A new wave of AI companies is racing to record how humans move, work, and interact with the world—potentially creating the training data that could power the next generation of robots.

WHAT’S HAPPENING

A growing number of AI companies are pursuing a new type of data collection that goes far beyond text, images, or videos.

Their goal is to capture human behavior itself.

The trend gained attention after a viral video appeared to show factory workers wearing head-mounted cameras while performing their jobs.

While the exact purpose of the footage remains unclear, the video sparked speculation that companies are gathering first-person human activity data to train future AI systems and robots.

One company operating in this space is Human Archive.

The startup recently raised $8.2 million from investors connected to major technology companies and AI ecosystems.

Human Archive equips workers with wearable cameras designed to capture how people:

• Move
• Work
• Interact with tools
• Navigate environments
• Perform physical tasks

The company says its objective is to build large-scale datasets focused on human sensimotor intelligence—the relationship between perception, movement, decision-making, and action.

Rather than simply teaching AI what humans know, the goal is increasingly to teach AI how humans operate in the physical world.

WHY IT MATTERS

Much of the AI revolution has focused on teaching machines to think.

The next phase may focus on teaching machines to act.

Large language models learned from internet-scale text.

Future robotic systems may learn from internet-scale human behavior.

That creates a potentially enormous shift.

For years, automation struggled with jobs requiring:

• Dexterity
• Adaptability
• Physical judgment
• Environmental awareness
• Hand-eye coordination

Collecting detailed first-person human data could help AI systems understand these activities at a much deeper level.

The result may be a new race to build datasets that capture not only human knowledge, but human experience itself.

The companies that own those datasets could gain significant advantages in robotics, automation, and embodied AI.

WHO BENEFITS

Robotics companies

Large behavioral datasets may accelerate the development of robots capable of performing more complex physical tasks.

AI infrastructure providers

New categories of data collection create demand for storage, processing, labeling, and model training systems.

Manufacturing automation firms

Improved robotic capabilities could expand automation opportunities across industrial environments.

Investors backing embodied AI

Companies developing physical-world intelligence may become increasingly attractive investment targets.

Researchers studying human intelligence

Large-scale behavioral datasets could provide new insights into how people learn, move, and interact with their surroundings.

WHO LOSES

Workers performing highly repetitive physical tasks

Jobs involving predictable manual labor may face increasing automation pressure.

Organizations without access to behavioral datasets

Data ownership may become a competitive advantage similar to what training data became for large language models.

Privacy advocates

The collection of detailed human activity data raises new concerns around consent, surveillance, and data ownership.

Companies slow to adopt automation

As embodied AI improves, competitive pressures may increase across industries.

Labor groups concerned about displacement

Some critics fear that today’s workers may be helping create the systems that eventually replace portions of their own workforce.

WHAT HAPPENS NEXT

The next major AI competition may not center on language models.

It may center on human behavior.

Companies around the world are likely to expand efforts to collect data involving:

• Physical movement
• Workplace tasks
• Tool usage
• Navigation
• Human decision-making
• Environmental interaction

The larger signal may be this:

The first AI boom taught machines how humans communicate.

The next AI boom may focus on teaching machines how humans operate in the real world.

And whoever builds the largest archive of human behavior could help shape the future of robotics and automation.

The Human Data Gold Rush Has Begun

Tagline

A new wave of AI companies is racing to record how humans move, work, and interact with the world—potentially creating the training data that could power the next generation of robots.

WHAT’S HAPPENING

A growing number of AI companies are pursuing a new type of data collection that goes far beyond text, images, or videos.

Their goal is to capture human behavior itself.

The trend gained attention after a viral video appeared to show factory workers wearing head-mounted cameras while performing their jobs.

While the exact purpose of the footage remains unclear, the video sparked speculation that companies are gathering first-person human activity data to train future AI systems and robots.

One company operating in this space is Human Archive.

The startup recently raised $8.2 million from investors connected to major technology companies and AI ecosystems.

Human Archive equips workers with wearable cameras designed to capture how people:

• Move
• Work
• Interact with tools
• Navigate environments
• Perform physical tasks

The company says its objective is to build large-scale datasets focused on human sensimotor intelligence—the relationship between perception, movement, decision-making, and action.

Rather than simply teaching AI what humans know, the goal is increasingly to teach AI how humans operate in the physical world.

WHY IT MATTERS

Much of the AI revolution has focused on teaching machines to think.

The next phase may focus on teaching machines to act.

Large language models learned from internet-scale text.

Future robotic systems may learn from internet-scale human behavior.

That creates a potentially enormous shift.

For years, automation struggled with jobs requiring:

• Dexterity
• Adaptability
• Physical judgment
• Environmental awareness
• Hand-eye coordination

Collecting detailed first-person human data could help AI systems understand these activities at a much deeper level.

The result may be a new race to build datasets that capture not only human knowledge, but human experience itself.

The companies that own those datasets could gain significant advantages in robotics, automation, and embodied AI.

WHO BENEFITS

Robotics companies

Large behavioral datasets may accelerate the development of robots capable of performing more complex physical tasks.

AI infrastructure providers

New categories of data collection create demand for storage, processing, labeling, and model training systems.

Manufacturing automation firms

Improved robotic capabilities could expand automation opportunities across industrial environments.

Investors backing embodied AI

Companies developing physical-world intelligence may become increasingly attractive investment targets.

Researchers studying human intelligence

Large-scale behavioral datasets could provide new insights into how people learn, move, and interact with their surroundings.

WHO LOSES

Workers performing highly repetitive physical tasks

Jobs involving predictable manual labor may face increasing automation pressure.

Organizations without access to behavioral datasets

Data ownership may become a competitive advantage similar to what training data became for large language models.

Privacy advocates

The collection of detailed human activity data raises new concerns around consent, surveillance, and data ownership.

Companies slow to adopt automation

As embodied AI improves, competitive pressures may increase across industries.

Labor groups concerned about displacement

Some critics fear that today’s workers may be helping create the systems that eventually replace portions of their own workforce.

WHAT HAPPENS NEXT

The next major AI competition may not center on language models.

It may center on human behavior.

Companies around the world are likely to expand efforts to collect data involving:

• Physical movement
• Workplace tasks
• Tool usage
• Navigation
• Human decision-making
• Environmental interaction

The larger signal may be this:

The first AI boom taught machines how humans communicate.

The next AI boom may focus on teaching machines how humans operate in the real world.

And whoever builds the largest archive of human behavior could help shape the future of robotics and automation.