RRReasoning Referee
← All discussions

MEDIA & CULTURE • AUTHENTICITY

When content stops feeling human

A REASONING REFEREE ORIGINAL · 2026

Both sides say “better communication.” They disagree about whether efficiency can scale without erasing authorship, trust, and meaning.

This is an illustrative exchange, not a report about a particular city, company, park, resident, employee, customer, driver, regulator, recipient, advocate, or business.

Woodcut illustration of a hand drawing one image beside a computer producing several similar images
CMO Chief marketing officer
“Our customers expect constant, relevant communication across every channel. Generative AI lets a small team produce more variations, answer questions faster, and spend human talent on strategy instead of repetitive production. Rejecting the technology because some content is bad would be like rejecting cameras because some photographs are bad.”
THE EXCHANGE
CD Creative director
“The problem is not that AI occasionally makes weak content. We are replacing human observation and judgment with statistically plausible material, publishing vastly more of it, and calling the result authentic. Efficiency means very little if customers stop believing that anyone is actually speaking to them.”

THE REASONING REFEREE

Make the call

Conditional rulingAI use is not automatically valuable or harmful. The call depends on the communication’s purpose, stakes, disclosure, accountability, customer value, and effects on human creative capability.
SCENARIO

A consumer brand is under pressure to publish more advertising, social posts, product descriptions, customer-service messages, and personalized content without substantially increasing its budget. Executives propose expanding generative AI throughout the operation. The resulting content is polished, frequent, and inexpensive, but customers begin calling it generic, artificial, and emotionally empty—“AI slop.” Creative employees warn that algorithmic volume is exhausting audiences and weakening trust. Executives respond that customers still expect convenience, personalization, and rapid service. The company must decide which communication should remain human-led, where AI may assist, what should be disclosed, and what success should mean.

A framework is a reasoning lens: the underlying question or principle someone uses to judge a decision.
A fallacy is a flaw in an argument. It can make a conclusion sound stronger than the reasoning actually supports.
A bias is a predictable mental shortcut that may influence what we notice, remember, or give extra weight.
A stake is something a person or group may gain, lose, protect, or be blamed for—even when it is not stated aloud.
01

Compared with what?

Opportunity Cost

AI can save production time and money, but the company must count any sacrifice in distinctive voice, trust, creative employment, accountability, and long-term brand identity.

Learn about Opportunity Cost→A real tradeoffTwo valuable uses of the same limited resource are competing. Choosing one means giving up some of the other.
02

Grow the pie, even if slices change

Kaldor–Hicks Improvement

AI adoption may create more total value while harming some workers or audiences. Efficiency claims remain incomplete until the company identifies who gains, who loses, and whether customer value survives.

Learn about Kaldor–Hicks Improvement→Potential improvementThe proposal may create more total benefit than harm, but that conclusion depends on credible evidence about who gains, who loses, and by how much.
03

Who decides, and who pays?

Principal–Agent Analysis

Executives may receive savings and greater output while creators face displacement and audiences sort through more synthetic material. Creative teams may protect their work while the company bears slower production costs.

Learn about Principal–Agent Analysis→Cost may be transferredThe decision-maker may receive the benefit while another person absorbs part of the burden or risk.
THE WIN-WIN CHECKCan this proposal be improved before anyone is asked to lose?

A better option to test

Keep original reporting, personal stories, sensitive service, artistic campaigns, apologies, and consequential claims human-led. Use AI as an accountable assistant for drafting, translation, accessibility, formatting, and low-risk operational messages—with disclosure when synthetic production changes how an audience would interpret the work.

THE PARETO TEST

Customers could receive faster, more accessible communication while human creators retain responsibility, authorship, and the work that requires lived experience. The test fails if automation misleads audiences, eliminates meaningful recourse to a person, or creates more attention cost than customer value.

Learn about Pareto Improvement

A DEEPER LOOK

Examine the strongest case on both sides.

THE QUESTION BEING TESTEDShould the brand expand its use of generative AI across customer service, advertising, and creative communication?

FOR • Expand responsible AI use

The executive’s strongest case

  • Many communication tasks are repetitive rather than deeply creative; faster drafting can reserve human time for judgment, research, interviews, and strategy.
  • AI can help smaller organizations compete, adapt material for different audiences, and provide quicker service without enormous content budgets.
  • Poor, derivative content existed before generative AI; quality depends on direction, editing, standards, disclosure, and human accountability.
AGAINST • Preserve human-led communication

The creative director’s strongest case

  • Near-zero production cost lets organizations flood attention channels with material that consumes more audience time than value it creates.
  • Synthetic communication can imitate insight or emotion without lived experience, investigation, authorship, or responsibility.
  • Undisclosed AI use can make audiences feel manipulated when a brand presents generated people, experiences, testimonials, or emotion as authentic.
REBUTTAL TO “FOR”

Producing more variations faster does not establish that audiences receive more meaning, usefulness, or trust. Cheap production can simply scale mediocrity.

REFEREE’S ANALYSISThis defeats volume as the primary success measure, but not carefully supervised AI use that improves a defined customer outcome.
REBUTTAL TO “AGAINST”

Human origin does not guarantee authenticity or quality. People also produce formulaic, deceptive, derivative, and low-value communication.

REFEREE’S ANALYSISThis defeats a blanket human-good, AI-bad distinction, but not the need for disclosure, authorship, and human responsibility where those qualities affect interpretation.
BEFORE A FIRMER RULING

Evidence that would matter most

  • Customer response to comparable content with and without clear AI disclosure
  • Trust, unsubscribe, blocking, complaint, retention, and conversion measures—not clicks alone
  • Quality, correction time, error rates, and true production costs for human, AI-assisted, and autonomously generated work
  • Effects on creative employment, skill development, brand distinctiveness, and customer access to a responsible human
RESEARCH USED TO FRAME THIS DISCUSSION

MOTIVATIONS & INCENTIVES

Why might they hold this position?

Consider good-faith reasons and possible competing interests side by side.

We begin by giving everyone the benefit of the doubt.

Our starting assumption is that people and institutions generally believe they are doing the right thing. Most advocates sincerely believe their position will lead to a better outcome.

At the same time, people can be influenced by incentives, responsibilities, experiences, and competing interests—sometimes consciously and sometimes unconsciously. The possibilities below are not accusations or conclusions. They are questions worth considering and testing against evidence.

BENEFIT OF THE DOUBT

Charitable Motivations

  1. Make useful communication affordableExample: Leaders may sincerely want a small team to serve more customers without an unaffordable production operation.
  2. Remove repetitive workExample: Executives may believe automation will free employees for research, judgment, strategy, and human relationships.
  3. Protect creative meaningExample: Artists and customers may genuinely want communication grounded in lived experience and accountable human choices.
  4. Preserve audience trustExample: Employees may worry that undisclosed synthetic communication will make every brand message seem suspicious.
  5. Improve accessibility and personalizationExample: AI supporters may want information adapted to different languages, abilities, knowledge levels, and customer needs.

POSSIBILITIES, NOT CLAIMS

Potential Hidden Agendas

  1. Labor-cost reductionExample: Language about innovation may partly conceal a plan to replace writers, artists, support representatives, or agencies.
  2. Professional self-preservationExample: Creative workers may oppose valuable automation partly because it threatens income, status, or control.
  3. Technology-vendor growthExample: AI companies benefit when ordinary production is redefined as impossible without their systems.
  4. Metric optimizationExample: Executives may favor clicks, posting frequency, and response speed because those are easier to measure than trust or cultural value.
  5. Avoiding accountabilityExample: An organization may blame an automated system for misleading or offensive communication that humans chose to deploy.

THE REFEREE’S OBSERVATION

More than one motivation can be true.

Someone can sincerely believe they are doing the right thing while also benefiting personally or institutionally if their preferred outcome succeeds. These explanations are not mutually exclusive. Look for evidence before assigning a motive.

Questions the Referee would ask

  • What evidence supports the charitable explanation?
  • What evidence supports an incentive-based explanation?
  • Could both explanations be true at the same time?
  • Would this person likely take the same position without a personal benefit?
  • What evidence would change my mind?

Reasoning Referee is not here to influence how you think. It is here to improve the quality of your thinking.

Rather than advocating for one viewpoint, it encourages intellectual curiosity, fairness, and critical evaluation—so the conclusion you reach is genuinely your own.

Good reasoning doesn’t guarantee the right answer—but it dramatically improves your chances of finding one.

REFEREE’S RULING

AI is not the same thing as AI slop—but cheap production makes slop easier to scale.

The executive is right that generative AI can remove repetitive work, expand accessibility, and help small teams communicate more effectively. The creative director is right that volume, speed, and short-term engagement do not establish meaning, authenticity, or durable trust. The strongest policy judges AI by the function and stakes of the communication. It preserves identifiable human responsibility, requires disclosure where synthetic production could affect interpretation, and measures whether the content creates value for the audience—not merely whether the company can create more of it.

THE QUESTION THAT MOVES IT FORWARDWhich parts of this communication require human experience, judgment, accountability, or relationship—and which can be automated without misleading the audience or reducing what it receives?