The digital transformation that made remote work possible also created enormous records of how people think, communicate, and solve problems. As companies turn that information into AI assets, a larger question emerges: Who benefits when the knowledge employees create becomes technology capable of performing some of their work?
THE SIGNAL
For decades, one of the most valuable assets inside a business was something that rarely appeared on its balance sheet: the accumulated knowledge of its employees.
Experienced workers knew how to solve unusual problems, handle difficult customers, navigate complex systems, and make decisions that kept organizations operating.
Much of that knowledge existed in conversations, personal experience, and professional relationships.
Then the workplace changed.
The pandemic accelerated a shift toward remote and hybrid work, moving everyday collaboration onto digital platforms. Meetings became video calls. Discussions became searchable messages. Project decisions became shared documents. Informal conversations increasingly left electronic records.
What began as a transformation in where people worked also changed how organizational knowledge was recorded.
Now artificial intelligence is creating a potential market for that information.
One remarkable example involves Google’s proposed $10 million purchase of business data from bankrupt Spirit Airlines.
Google won an August 2026 auction for data that included internal communications, spreadsheets, calendars, operational information, and productivity records. The company said it intended to use the material for product development and AI training.
The transaction has attracted scrutiny over employee privacy and the commercial use of workplace records.
On November 10, Google agreed to stronger protections, including additional exclusions of sensitive employee information and independent verification of de-identification. Judicial approval remained pending.
The broader implications extend beyond a single airline.
Information generated through ordinary work is becoming a potentially valuable raw material for AI development.
That could transform the relationship between employees, employers, and technology companies.
The question is no longer simply whether artificial intelligence will replace jobs.
It is whether the knowledge created by workers can become a separate corporate asset—and eventually help automate portions of the work those employees once performed.
WHAT THE MARKET IS MISSING
Most discussions about AI and employment focus on productivity, labor costs, and job displacement.
But there is another transformation taking place beneath those debates.
Companies are beginning to recognize that the records of how work gets done may have substantial value beyond their original business purpose.
Consider what an organization accumulates over ten years.
Millions of emails explain decisions, resolve problems, and coordinate projects. Customer-service conversations reveal common complaints and successful responses. Internal documents describe procedures. Software repositories preserve technical solutions. Meeting transcripts record negotiations, explanations, and institutional decisions.
Together, these materials can form a detailed record of organizational activity.
AI systems may be able to extract patterns from this information, organize it, retrieve relevant knowledge, and assist with tasks that once required employees to recall or interpret past experiences.
There is an important distinction, however.
Not every digital communication is automatically used to train AI models. Many enterprise systems use retrieval methods that allow AI to consult authorized documents without incorporating them into a model’s underlying training data.
Training, searching, summarizing, and automating are different activities, with different privacy and commercial implications.
Yet all reflect a larger development: workplace information is becoming more accessible to automated systems.
The Spirit Airlines example
The proposed Google transaction reveals the potential scale of this opportunity.
According to lawmakers examining the sale, the data originally under consideration could include approximately 100 million emails and 500 million Microsoft Teams messages or collaboration records. Those figures describe the scope cited in scrutiny of the transaction, not necessarily what Google will ultimately receive.
Google has said it does not seek personal information. Additional safeguards negotiated with unions are intended to exclude sensitive material and prevent individuals from being re-identified.
But consider the economic significance.
A company that ceased operations in May could still possess business information valuable enough to attract a multimillion-dollar bid.
Its physical operations had stopped.
Its accumulated digital knowledge retained commercial value.
The overlooked signal is that an organization’s history of work may become an asset that survives the organization itself.
For employees, this introduces an unfamiliar economic question.
They were compensated for performing their jobs. Their communications and decisions helped create the records that the company retained.
If those records later contribute to valuable AI products, should workers receive any additional benefit?
There is no universal answer.
Employment agreements, intellectual-property rules, confidentiality obligations, privacy laws, and collective-bargaining arrangements can affect the rights involved.
But the question becomes increasingly important as organizations discover new ways to commercialize digital knowledge.
FIRST-ORDER EFFECTS
The most immediate consequences involve the value of corporate data, the capabilities of workplace AI, and growing scrutiny over how employee information is used.
Corporate records become AI resources
Businesses have traditionally preserved information for compliance, operational continuity, customer service, and management.
AI adds another possible use.
Existing records can help organizations develop systems that answer internal questions, locate relevant documents, recognize recurring problems, and support decisions.
This may reduce the time employees spend searching for information or repeating routine procedures.
A new employee could potentially learn an organization’s practices through an AI assistant connected to approved internal materials.
A customer-service representative could receive suggestions based on successful previous interactions.
The value arises not simply from collecting more information, but from making useful knowledge available when needed.
Experience can be distributed more broadly
Research offers evidence that AI can help transfer some benefits of accumulated workplace experience.
A study published in The Quarterly Journal of Economics in 2025 examined 5,172 customer-support workers. Access to an AI assistant increased productivity by approximately 15% on average.
Less experienced workers experienced particularly meaningful improvements, while the most experienced workers saw comparatively small benefits.
The researchers found evidence consistent with AI helping spread effective practices among workers.
This suggests an important possibility.
Knowledge that once gave experienced employees an individual advantage may become easier to share through technology.
That can benefit organizations and less experienced colleagues.
But it also raises questions about how the economic value of experience will be rewarded if software can increasingly reproduce some of its practical benefits.
Employee privacy becomes a business issue
Workplace communications can contain more than business instructions.
They may reveal personal circumstances, interpersonal relationships, performance concerns, confidential discussions, and sensitive information about individuals.
Removing names does not automatically eliminate every privacy risk.
Information from several sources can sometimes be combined to reveal identities or sensitive patterns.
The controversy surrounding the Spirit Airlines transaction illustrates why protections matter.
Labor unions objected to the proposed sale and sought stronger restrictions. The November agreement provides for independent review of de-identification, exclusions of sensitive employee records, confidentiality requirements, and limits on Google’s use of identifiable information.
These protections also demonstrate that employee information is not simply an unrestricted commodity.
The terms under which workplace data can be transferred and used are increasingly part of commercial negotiations.
SECOND-ORDER EFFECTS
The deeper consequences could emerge as businesses move beyond using AI to assist employees and begin redesigning work around automated systems.
A new market for institutional knowledge
The Spirit Airlines transaction suggests that organizations could increasingly view operational records as assets with independent commercial value.
A company’s communications, processes, and business history might help train or evaluate AI systems even after the company stops operating.
That creates opportunities for technology developers, data providers, and organizations holding extensive records.
It also creates complications.
A company’s information may contain intellectual property belonging to customers, suppliers, software vendors, or other third parties.
Corporate ownership of a document does not necessarily grant unrestricted rights to use every piece of information within it.
Future transactions may therefore require more detailed rules governing data selection, privacy protection, intellectual property, and permitted AI applications.
The possibility of employees training their own replacements
This is the most consequential—and most uncertain—part of the story.
Suppose a business uses years of employee-generated records to develop an AI system that handles routine customer questions, prepares standard reports, or coordinates administrative workflows.
Some responsibilities that previously required employees might become automated.
The information created during years of human work would have helped make that possible.
But the evidence does not establish that this process will eliminate entire occupations.
An AI system capable of summarizing successful customer interactions may still struggle with unusual situations, conflicting instructions, accountability, or decisions requiring human discretion.
Businesses may also use productivity gains to expand services rather than reduce employment.
The potential for substitution is real, but the outcome depends on the technology, the work involved, and management decisions.
The structural risk is that knowledge created collectively by workers could be used to reduce an organization’s future dependence on some of those same workers.
Experience could lose some of its scarcity value
Companies have traditionally relied on experienced employees because those workers possess knowledge that is difficult to transfer.
They know which processes work, which problems recur, and how to respond when circumstances become complicated.
AI systems connected to institutional knowledge could make portions of that experience accessible to other employees.
This may improve productivity and reduce dependence on particular individuals.
It could also change the bargaining power associated with experience.
If an organization can preserve and retrieve valuable operational knowledge independently of the employee who developed it, some forms of individual expertise may become less scarce.
However, experience includes judgment, relationships, intuition, and accountability that are not fully captured in digital records.
The likely distinction is between knowledge that can be documented and reused, and expertise that must still be exercised by people.
The distribution of AI wealth becomes a larger issue
If employee-generated information contributes to profitable AI systems, who should receive the economic rewards?
Companies may argue that records created during paid employment belong to the organization and can legitimately support business innovation.
Workers and labor representatives may argue that new commercial uses—particularly those affecting privacy or employment—justify additional protections, consultation, or negotiated benefits.
Technology companies may emphasize the investment required to transform raw information into functioning AI systems.
These are competing economic interests, not questions with simple answers.
Possible arrangements include collective bargaining over data use, restrictions on secondary commercialization, profit-sharing agreements, and stronger requirements for transparency.
None is an inevitable outcome.
But the growing commercial value of workplace records may force organizations to address questions that were less urgent when those records served primarily operational purposes.
WINNERS
AI developers could benefit from access to specialized business information that helps improve systems intended for professional use. High-quality records may provide examples of real organizational processes that are difficult to reproduce through artificial data alone.
Businesses with organized, accessible information may gain an advantage over competitors whose knowledge remains fragmented across incompatible systems or undocumented practices.
Small businesses could benefit from AI systems that make sophisticated operational knowledge and automated assistance more affordable.
New and less experienced employees may gain faster access to useful information, established procedures, and examples of successful work.
Experienced professionals who manage AI effectively could become more valuable by helping organizations determine which knowledge can be automated, which systems are reliable, and where human judgment remains essential.
Data-governance and security providers may also see growing demand as companies seek to manage the risks associated with using confidential information in AI systems.
The common advantage is not merely possessing large quantities of data.
It is having useful information, legitimate access to it, and the capability to turn it into reliable outcomes.
LOSERS
Workers whose responsibilities involve highly repeatable digital processes could face greater automation pressure as AI systems become better at handling standardized work.
Employees with limited visibility into corporate data practices may have little understanding of how their communications are retained, analyzed, or used for purposes beyond their original tasks.
Professionals whose bargaining power depends on exclusive access to institutional knowledge may encounter pressure as AI makes some documented expertise more accessible.
Businesses with weak information-security practices risk exposing sensitive material through poorly controlled AI integrations, inappropriate data sharing, or insufficient access restrictions.
Organizations that fail to build employee trust may experience resistance to AI adoption if workers believe their information is being used without appropriate protections.
However, workers are not automatically disadvantaged when companies adopt AI.
Technology that reduces repetitive work can improve job quality, enable employees to take on more valuable responsibilities, and support business growth.
The distribution of benefits will depend partly on how organizations implement AI and whether productivity gains translate into better wages, working conditions, opportunities, or employment.
WHAT HAPPENS NEXT
The immediate test will be whether Google’s proposed Spirit Airlines data acquisition receives judicial approval and under what conditions.
As of November 13, 2026, a bankruptcy court hearing is expected the following week, after unions negotiated additional employee privacy protections.
The outcome could provide an important example of how companies, employees, and courts address the transfer of business information for AI development.
Beyond that transaction, five developments deserve attention.
First, corporate data inventories may become more valuable. Businesses could begin evaluating internal communications, process documentation, and operational records as potential AI resources rather than merely historical archives.
Second, employee agreements may evolve. Employment contracts, collective-bargaining arrangements, and internal policies could increasingly address AI training, data retention, secondary uses, and employee privacy.
Third, demand for verified institutional knowledge could increase. Organizations may invest in better documentation and information management so AI systems can retrieve accurate, current records instead of relying on fragmented or obsolete material.
Fourth, automation may advance unevenly. Routine digital workflows are attractive candidates, but their automation will depend on technical reliability, cost, regulatory obligations, and the need for human intervention.
Fifth, the economic benefits may become a workplace negotiation issue. If corporate records contribute materially to AI productivity or commercial products, questions about employee participation in those benefits could become more prominent.
The most important indicator will be whether companies move from using AI to assist workers toward systematically using accumulated workplace knowledge to replace particular responsibilities.
That would signal a deeper change in the relationship between human experience and organizational technology.
BOTTOM LINE
Remote work did more than change where employees sit.
It accelerated a broader movement toward digital collaboration, making more workplace activity recordable, searchable, and potentially reusable.
Artificial intelligence is giving that information new economic significance.
A conversation that once helped resolve a problem may now become part of an organizational knowledge repository.
A documented procedure may support an AI assistant.
A collection of business records may attract buyers seeking information to develop new technology.
These developments can improve productivity, preserve valuable expertise, and help employees perform their jobs more effectively.
But they also create an important imbalance.
Employees contribute knowledge through their work. Employers frequently control the systems where that knowledge is recorded. AI developers may possess the technology needed to extract additional commercial value from it.
Whether workers share in those benefits—and how their privacy and professional interests are protected—remains an open question.
The proposed Google–Spirit Airlines transaction is significant because it makes this transformation tangible.
An airline stopped operating, yet its accumulated records remained valuable enough to attract a multimillion-dollar AI-related bid.
That does not prove employees are being replaced by systems trained on their own work.
It establishes something more immediate: business information that once supported an organization’s daily operations can become a marketable resource for AI development.
The larger signal: We may be entering an economy in which the knowledge people create at work becomes increasingly valuable independently of the people who created it.
The challenge is not simply determining what artificial intelligence can automate.
It is deciding how the value created from human knowledge should be controlled, protected, and distributed.
The next great workplace debate may not be about where people work.
It may be about who owns the lasting value of what they know.