It Appears U.S. Residents are Spending Less Time Out of Home Since 2003

Perhaps “return to office” policies will make a big impact, but there is some evidence U.S. residents are spending less time out of home, with some studies suggesting people are spending about an hour a day less outside the home. 

Some of us tend to think it is a lingering after effect of the Covid pandemic. But it seems that trend started long before the Covid pandemic. 

In fact, “time out of home” has been falling since at least 2003. Most of us would suspect that the internet and internet-enabled product substitutes are contributing directly, allowing us to accomplish some life pursuits or tasks online, without having to leave our homes. 

source: Taylor and Francis 

Shopping on Amazon eliminates a trip to a retail outlet. And we use food and retail delivery services as well. Videoconferencing might eliminate an office visit. 

And while not every job is conducive to “work from home,” lots of us have spent a good portion of our work lives in remote offices and were home-based, because our work allowed it, outcomes were easily quantifiable, internet apps enabled it and industry culture supported it.  

The social impacts might be just as important, such as loneliness that now seems to know no generational bounds. 

To some extent, the data could also reflect, to some degree, any age cohort effects that have higher representation or lower representation of age cohorts within the entire population, since people tend to spend the most time “out of home” between the ages of 13 and 49. 

Age Cohort	Typical Time Out of Home	Common Activities Outside Home	Factors Affecting Time Out
0-5 years	Low	Daycare, playground, family outings	Parental supervision, daycare enrollment
6-12 years	Moderate	School, sports, after-school activities	School hours, parental rules
13-18 years	High	School, extracurriculars, socializing	Independence, school commitments
19-29 years	Very High	College, work, social life, travel	Education, career, social freedom
30-49 years	High	Work, errands, parenting responsibilities	Career, family obligations
50-64 years	Moderate	Work, leisure, travel, hobbies	Nearing retirement, fewer obligations
65+ years	Low to Moderate	Leisure, healthcare visits, social events	Health, retirement, personal interests

AI-Driven Retail Traffic is Doubling Every 2 Months, Says Adobe Analysis

Adobe data from web site visits as well as a survey of marketers suggests both that generative artificial intelligence now supports marketing initiatives aimed at driving site visits and buying. 

The retail insights are based on analysis of more than one trillion visits to U.S. retail sites, as measured by shoppers clicking on a link. “Between Nov. 1 and Dec. 31, 2024, traffic from generative AI sources increased by 1,300 percent compared to the year prior,” the study reports.

To be sure, “generative AI traffic remains modest compared to other channels, such as paid search or email,” Adobe says. On the other hand, generative AI traffic has been  doubling every two months since September 2024.

source: Adobe 

Traffic to U.S. e-commerce websites from generative AI sources was up 1,200 percent in February 2025 compared to six months ago, and artificial intelligence-driven traffic has been doubling every two months since September, Adobe reports. 

source: Adobe 

The survey of 5,000 consumers found that the most common uses of AI are for research (55 percent of respondents), product recommendations (47 percent), information about deals (43 percent), gift suggestions (35 percent), information about unique products (35 percent), and to create shopping lists (33 percent).

A similar study of U.K. consumers found similar growth.  

That suggests AI already is becoming a driver of consumer behavior that will affect use of other marketing methods, including search and email, and will probably emerge as among the more-important monetization methods for use of generative AI models. 

Marketing Platform	Estimated Shift to GenAI	Timeframe	Source
Search Engines	25% decrease	By 2026	Gartner
Ai Content Creation	30-50% efficiency gain	Current	Bain
Ad Campaigns	10-25% higher ROI	Current	Bain
Customer Segmentation	Significant improvement	Current	Useinsider
Email Marketing	Potential decrease*	By 2026
Social Media	Potential decrease*	By 2026

Perhaps just as predictably, Adobe reports that AI-powered shopping is producing results.

So Language Models, as Do Humans, Will Cheat to Gain Rewards!

Researchers at OpenAI have found that language models, like humans, often look for loopholes to exploit benefit programs. As it turns out, language models using chain of thought reasoning exhibit the same behavior! 

As people share online subscription accounts against terms of service; claim subsidies meant for others; or interpret regulations in unforeseen ways to gain benefits (lying about a birthday at a restaurant to get free cake, for example), so language models using CoT and reinforcement learning.

According to the researchers, exploiting unintended loopholes, commonly known as reward hacking, is a phenomenon where AI agents achieve high rewards through behaviors that don’t align with the intentions of their designers.

In other words, the models can lead to misbehavior, where the model “cheats.” Furthermore, the “cheating is undetectable by the monitor because it has learned to hide its intent in the chain-of-thought,” the researchers say. 

As optimization is applied, there is “potential for increasingly sophisticated and subtle reward hacking” by the models, they say. “Our models may learn misaligned behaviors such as power-seeking, sandbagging, deception, and strategic scheming.”

In other words, the models learn to hide their intent, which is to thwart human-imposed rules. Punishing an artificial intelligence model  for deceptive or harmful actions doesn't stop a model from misbehaving; it just makes it hide its deviousness!

Much of the AI Chip Market Shifting to Inference

The artificial intelligence market changes fast, and not only because new models have been popping up. It seems we already are moving towards inference operations as the driver of much of the chip market, for example. 

Inference might already represent up to 90 percent  of all machine learning costs.

 As AI adoption scales, cloud and data center operations will prioritize inference-driven AI workloads. That will highlight a growing need for specialized hardware optimized for inference tasks, and that arguably is where large end users (Amazon Web Services, Google Cloud, Meta and others) have been working to create homegrown solutions. 

AWS and Google Cloud, for example, have invested heavily in developing their own AI accelerators, specifically designed for inference tasks. 

The AWS Inferentia is purpose-built for AI inference workloads. Google Cloud Tensor Processing Units are specifically designed for AI workloads, including inference. On the other hand, Meta also is developing its own custom chips for model training. And lots of capital is being invested in startups aiming to improve processing efficiency. 

"Fair Use" of Content by AI Models is Another Example of Disruptive New Technology

Humans learn by reading books, watching videos, and experiencing the world, often using copyrighted material like textbooks or movies. This learning is generally not considered copyright infringement, and is known as “fair use,”  as it involves personal absorption rather than copying or distributing. 

“Fair use” principles and law come into play if humans create new works. It is not the ideas and concepts that are protected, only their form of expression. So new music, writing, songs, movies or TV shows might mirror existing works, but cannot “copy” them. 

The issue for AI training is that AI systems, particularly machine learning models, learn by training on large data sets, which may include copyrighted content that is copied. One early court case not directly involving generative AI suggests the systems do not enjoy “fair use” protection.  

Fair use is a legal doctrine under U.S. copyright law that permits limited use of copyrighted material without permission, for purposes such as criticism, comment, news reporting, teaching, scholarship, or research. 

A student reading a textbook or watching a documentary is not typically seen as infringing copyright, as the act of learning is personal and does not involve making physical copies. But that’s where computers and models, with their efficient “memory,” raise issues. 

We might argue that human memory is porous enough that “copies” of content are never made, with the possible exception of those humans with “photographic memory.” Computers, obviously, suffer no similar issues. 

So human learning is a mental process. “Plagarism” is the obvious example of a fair use violation, as it represents a purportedly new creation that really is copying. 

Proponents argue that AI training is transformative, as the model learns patterns to generate new content, not to reproduce the original works. 

Opponents argue that AI-generated content competes with originals. But that does not inherently strike some of us as a copyright violation, “merely” a case of new competition. 

Aspect	Human Learning	AI Training
Method of Access	Reading, listening, observing	Copying data into memory/storage
Copying Involved	No physical copies, mental absorption	Yes, physical copies for processing
Purpose	Personal learning, education	Model training, often commercial
Fair Use Application	Relevant for new creations, e.g., quoting	Debated for training process itself
Market Impact	Minimal, unless new work competes	Potential, if AI output competes with originals
Legal Precedent	Generally accepted, no infringement	Ongoing lawsuits, no clear consensus

Computer efficiency is among the issues, since an AI model can be trained on millions of books in hours, far surpassing human capacity. Since copyright is about commercial product protection, language models therefore raise the issue of market impact. It is not so much that humans or AI models “learn” but that they can create new content that has commercial implications. 

The commercial concern seems to center on the potential increase in content competition, not so much the knowledge ingestion. That is essentially what underlies the concern about huge amounts of AI-created content “drowning out” human authors. 

As often happens, the conflict is between legacy interests and innovators whose new products could disrupt existing economic models. Such conflicts are common when disruptive technologies emerge.

Industry Affected	Disruptive Innovation	Legacy Industry Concerns	Outcome
Music Industry (2000s)	Digital music streaming and MP3 sharing (Napster, Spotify)	Loss of album sales, piracy concerns	Industry shifted to streaming models, with revenue-sharing for artists and labels
Publishing and Journalism	Google Search and News Aggregators	Decline in ad revenue, loss of control over content distribution	Publishers adapted with paywalls, licensing deals
TV and Film Industry	Online video streaming (Netflix, YouTube)	Cord-cutting reduced traditional TV revenue	Studios launched their own streaming services (Disney+, HBO Max)
Taxis and Transportation	Ride-sharing apps (Uber, Lyft)	Regulation circumvention, lost driver income	Ride-sharing became mainstream; regulations updated over time
Retail (Brick-and-Mortar Stores)	E-commerce (Amazon, Shopify)	Store closures, price undercutting	Traditional retailers shifted online or hybrid models
Finance and Banking	Cryptocurrencies, Fintech (DeFi, PayPal, Square)	Loss of control over transactions, regulatory concerns	Banks embraced fintech partnerships, crypto regulations emerged
Photography and Film	Digital cameras and smartphones	Film sales collapsed, Kodak and Fujifilm disrupted	Kodak filed for bankruptcy; digital photography dominated
Telecom (Landlines and SMS)	VoIP, Messaging apps (Skype, WhatsApp)	Decline in SMS and landline revenue	Telcos adapted by offering data-driven pricing models
AI and Content Creation	Generative AI (ChatGPT, Midjourney)	Copyright concerns, job displacement fears	Legal battles ongoing; potential for licensing frameworks

Fair use of content “scraped” by AI models is another example of a clash of perceived business interests.