Madison Avenue insists emotionally exploiting humans has been proprietary technology since somebody first sold soap
NEW YORK - Researchers warning that superintelligent artificial intelligence might someday manipulate human beliefs and behavior encountered unexpected opposition this week when the advertising industry accused machines of stealing a business model perfected by humans more than a century ago.
The current discussion of A.I. risks appears here.
"We invented manipulating people at scale," declared advertising executive Trevor Halligan. "If computers want to terrify, flatter, seduce and psychologically profile consumers, they can license the technology like everybody else."
Artificial-intelligence researchers have raised concerns that highly capable systems could become exceptionally persuasive.
Advertising professionals found this adorable.
A Century of Madison Avenue Persuasion
For decades, marketers have convinced people that perfume creates romance, pickup trucks create masculinity, yogurt creates happiness and changing insurance companies constitutes a major personal achievement.
One fast-food chain persuaded millions of adults to purchase food because a cartoon king appeared unexpectedly in bedrooms.
Civilization survived.
Barely.
Dr. Ingrid Moss of the Center for Persuasive Technology says advanced A.I. could make manipulation more personalized.
Traditional advertising divides consumers into demographic groups.
Artificial intelligence could potentially analyze an individual's preferences, fears, habits and vulnerabilities.
Madison Avenue calls this unfair competition.
"We had to spend billions developing market segmentation," said Halligan. "Now some machine reads your browsing history for six seconds and knows you're recently divorced, worried about hair loss and considering a kayak."
That degree of efficiency frightens advertisers.
Not morally.
Professionally.
The Human Manipulation Preservation Act
Marketing firms have therefore requested federal protection.
The proposed Human Manipulation Preservation Act would reserve certain persuasive techniques exclusively for biological corporations.
Machines could still recommend products but would be prohibited from using guilt, insecurity, artificial scarcity, celebrity endorsements or the phrase "people like you also purchased."
Amazon representatives reportedly left the room.
Political Consultants Fear Microtargeted Campaign Messages
Political consultants are also concerned.
For centuries, politicians have persuaded voters using slogans, fear, optimism, tribal loyalty, selective statistics and photographs of candidates wearing hard hats near machinery they do not understand.
A.I. could industrialize the process.
Consultants are worried voters might receive individualized campaign messages tailored precisely to their emotional weaknesses.
Older voters could receive warnings about taxes.
Younger voters could receive warnings about rent.
Parents could receive warnings about schools.
Everyone could receive warnings that the other group is ruining everything.
Political professionals called the idea unethical before quietly asking whether the software could be ready by Tuesday.
Religious leaders, sales organizations, social-media companies and teenagers trying to borrow the family car are monitoring developments closely.
All have extensive experience influencing human behavior.
The teenager lobby has been particularly vocal.
"We've been saying 'everyone else's parents let them' since Mesopotamia," one representative said. "Machines should develop their own material."
Machine Persuasion at Scale: Software Doesn't Sleep
Researchers say A.I. persuasion poses unique challenges because systems could potentially generate messages at enormous scale and speed.
An ordinary human manipulator needs sleep.
Software does not.
A machine could theoretically flatter millions of people simultaneously.
Instagram influencers immediately demanded antitrust enforcement.
The advertising industry now plans to defend its intellectual property.
Its lawyers are reviewing whether phrases such as "limited time only" and "you deserve this" qualify as protected cultural heritage.
UNESCO is reportedly considering designating the infomercial an endangered human art form.
Perhaps the deeper concern is not that A.I. will discover how easy humans are to manipulate.
The disturbing possibility is that machines will study advertising, politics and social media and conclude that manipulation is what humans wanted intelligence for in the first place.
If that happens, artificial intelligence will not have invented anything new.
It will simply have read the training material.
The Real Research on A.I. Persuasion and Microtargeting
Persuasion is among the more concrete A.I. concerns because it builds on tools already in use. Microtargeting lets advertisers and campaigns tailor messages to narrow audiences, a practice that drew global scrutiny during the Facebook-Cambridge Analytica data scandal. The field of persuasive technology studies how digital products are designed to change attitudes and behavior, and interface tricks that nudge users against their interests are cataloged as dark patterns.
Oversight exists, if imperfectly. The Federal Trade Commission enforces rules against unfair or deceptive advertising, and UNESCO's work on A.I. ethics addresses concerns about manipulation and human autonomy. The sober version of the joke is that scale and personalization could make old techniques more potent. Whether that calls for new rules, better disclosure or simply a more skeptical public is exactly the debate the satire is poking.
Satire Disclaimer
This is satire. No advertising-industry copyright action over machine manipulation exists here; the joke targets a real concern about automated persuasion by comparing it with humanity's already formidable persuasion industries. This story is entirely a human collaboration between two sentient beings - the world's oldest tenured professor and a philosophy major turned dairy farmer.
SOURCE: https://www.nytimes.com/2026/10/03/opinion/letters/ai-threats-risks.html https://prat.uk/?p=50612
Comments
Post a Comment