The Problem With AI: Bullshit Has Never Looked This Good

There has always been bullshit on the internet. That isn’t new. What is new is how good it can look. For most of the web’s history, making yourself look like a serious business, an established company or somebody worth listening to involved a certain amount of friction. You needed a decent website, a recognizable identity, photographs, competent writing and, ideally, something resembling a track record. None of that guaranteed legitimacy, of course. The internet has never been a bullshit-free environment. But producing a convincing professional presence took time, money and expertise, and those requirements created at least some relationship between presentation and reality. AI is breaking that relationship. Today, somebody with little experience in an industry can use AI to create the name, branding, website, copy, imagery, founder biography, social posts and articles of an apparently established expert business. The result can look fantastic, sound authoritative and give every impression of competence. Underneath it, there may be almost nothing.

That is the part I think we are only beginning to understand. AI isn’t merely making it easier to produce content. It is making it dramatically cheaper to manufacture the outward signs of experience, authority and success. Somebody who has never completed a construction project can create the website of a construction consultant. Somebody with no meaningful history in an industry can publish a library of articles explaining its most complicated problems. A new company can present itself with the polish of an established organization, complete with professional-looking photography, convincing case studies and a carefully constructed narrative about its expertise. The question is no longer simply whether the information is well written or the website looks professional. The question is whether any of it represents something real.

Anybody Can Look Like Somebody

There is nothing inherently wrong with helping a new business look professional. Good branding and web design have always helped businesses communicate their value, and a small company shouldn’t have to look amateurish simply because it is small or new. In fact, one of the positive consequences of AI is that a legitimate business can now communicate more effectively without needing the budget of a much larger competitor. The problem begins when the presentation becomes disconnected from the underlying reality. Imagine somebody starting a consulting company tomorrow with limited industry experience, no substantial client history and little evidence of having solved the problems they claim to understand. They can still produce a beautiful website, an impressive About page, a convincing professional biography and twenty articles explaining the industry’s most important issues. They can create a visual identity that suggests maturity and authority, then use AI to maintain the appearance of an active, knowledgeable organization. To a visitor encountering the company for the first time, everything may look perfectly legitimate. That visitor has no immediate way of knowing whether the expertise is genuine or whether the company has simply become very good at describing expertise it doesn’t possess.

This is a different problem from the familiar complaint that AI will flood the internet with mediocre articles. The web has always had mediocre articles, inflated claims and people who talk a better game than they play. What AI changes is the scale, speed and cost of producing an entire identity around those claims. It can help manufacture not just a piece of information but the appearance of an organization: its voice, history, personality, expertise and professional presence. The individual elements can reinforce one another until the whole thing feels convincing. We are moving into a period in which the ability to look credible is becoming widely available, regardless of whether the underlying credibility has been earned.

Search Has Always Been in the Trust Business

People often talk about search engines as though their primary job is to find information. That is only part of the job. The harder problem has always been deciding which information deserves attention and which sources are most likely to satisfy the person searching. Google cannot personally inspect every business, interview every expert or verify every claim made on every website. It has to infer credibility from signals: links, mentions, reviews, history, content, relationships between websites, location, demonstrated experience and many other clues. No single signal proves that a business is trustworthy, and none is immune to manipulation. Together, however, they can help form a picture of an entity and its place in the world. This is one reason search has always been an evolving discipline. As people learn to manipulate a signal, search systems have to become more sophisticated about interpreting it.

AI raises the stakes because it makes some of the familiar signals of professionalism and authority much easier to manufacture. A polished website is useful, but it is no longer especially difficult to produce one. A large body of content can demonstrate knowledge, but it can also be generated without firsthand knowledge. Professional photography can establish a visual impression, but generated imagery can create that impression without documenting a real place, person or project. None of this means these signals have become worthless. It means they need to be understood in context rather than accepted at face value. The challenge for search is not to eliminate AI from the process. It is to distinguish useful work assisted by AI from manufactured authority that has little substance behind it.

That distinction connects directly to the argument in our discussion of changing Google rankings and search volatility. Search is not a static system in which a business discovers the right combination of words and technical adjustments, locks in a position and assumes the job is finished. Algorithms evolve because the web evolves, user expectations change and the incentives surrounding visibility change with them. The increasing availability of convincing synthetic content creates another reason for search systems to improve how they evaluate sources, relationships and evidence. The important question is not whether a particular page was produced with AI. It is whether the page, the entity behind it and the claims it makes deserve the confidence a search engine or an AI assistant places in them.

Information Is Cheap. Evidence Isn’t.

This is where the distinction between information and evidence becomes important. AI can explain commercial construction, describe search engine optimization, outline the principles of website architecture or produce a detailed discussion of almost any professional subject. That ability can be enormously useful. It can help people understand unfamiliar subjects, organize their thinking and communicate knowledge they might otherwise struggle to express. But the existence of a convincing explanation does not establish that the person publishing it has firsthand experience. A genuine expert can use AI to help communicate twenty years of experience, while somebody who has never done the work can use the same technology to produce an equally polished explanation. The words alone may not tell us which is which. The difference exists somewhere beyond the words.

Evidence is what connects a claim to something outside the claim itself. A contractor has completed projects. A photographer has documented actual assignments. A law firm has represented clients. A manufacturer has products in the world. A web agency has websites it has built. A consultant has people who can describe the work and its results. Businesses accumulate relationships, history, customer experiences, references and consequences from the decisions they make. These things do not automatically prove that a company is excellent, and genuine businesses can certainly exaggerate their accomplishments. But real work creates opportunities for verification that a carefully written paragraph cannot provide by itself. The more easily information can be generated, the more important it becomes to understand what supports it.

That is also why the ability to create website content with ChatGPT should not be confused with the ability to create expertise. AI can help turn experience into useful explanations, make complicated ideas easier to understand and help a small business communicate more consistently. Those are real advantages. But the tool does not magically supply the underlying experience simply because it can describe it fluently. The opportunity is to use AI to communicate what a business actually knows and has done, not to use it to conceal the absence of those things. Information can be generated on demand. A meaningful record of work has to be built.

The Old Signals Are Going to Get Cheaper

Think about how many things that once required specialized resources can now be produced with relatively little effort. Professional-looking imagery, polished copy, brand concepts, explanatory videos, presentations, diagrams and entire websites are increasingly accessible to people who would previously have needed several specialists to create them. This is good news for businesses that have something valuable to offer but lack the budget to communicate it properly. It is also good news for people who have ideas and need help expressing them. But there is a consequence: when a particular signal becomes cheap and widely available, its ability to distinguish one source from another can diminish. If nearly everybody can have an impressive website, the website’s appearance tells us less about the quality of the business behind it. If nearly everybody can publish thoughtful-sounding articles, the volume of published content becomes a weaker indicator of expertise. Presentation still matters, but presentation alone becomes less decisive.

That does not mean search engines will simply abandon traditional signals or that every business will need a mountain of external citations to be considered credible. Search systems are complicated, and their methods are not fully visible from the outside. It does mean there is a compelling incentive to evaluate signals in combination and to distinguish self-reported claims from information supported elsewhere. A business saying it is an industry leader is one thing. A business with identifiable projects, customers, relevant experience and independent recognition offers a richer set of clues. A page that explains a subject is useful. A page that demonstrates original knowledge, makes a defensible argument and connects to a recognizable body of work can offer more. The direction is not mysterious: when appearances become easier to manufacture, the ability to establish context and corroboration becomes more valuable.

Search Will Have to Look Beyond the Website

A company’s website remains central to its digital presence, and I would argue that a good website becomes more important rather than less important in this environment. It is where a business explains who it is, what it does, how it works and why somebody should consider hiring it. It can document projects, organize expertise, provide useful resources and make the company’s identity understandable to both people and machines. But the website is also a place where the company controls the presentation. It chooses the claims, the photographs, the case studies and the testimonials it displays. That is not a flaw; it is how marketing works. It simply means that the website is one source of information about the business, not independent proof of every claim the business makes.

The wider web can provide additional context. Do other organizations mention the company? Do customers describe experiences that align with its claims? Does its work appear in relevant places beyond its own marketing? Is there a consistent history of activity? Are the people associated with the business connected to the expertise they claim? Are the company’s public claims supported by identifiable work? None of these questions can be answered reliably by a single universal test, and independent mentions are not automatically trustworthy. But together they illustrate why entity recognition, reputation and corroboration matter. The task is to build a more complete picture than any one website can provide. This is especially important as AI assistants increasingly synthesize information and make recommendations rather than simply displaying a list of pages for users to evaluate themselves.

A search result can give a user ten possibilities and leave the evaluation to them. A recommendation system is taking on a different responsibility: it is attempting to identify which possibility is a good answer to a particular need. That does not mean AI recommendations are necessarily accurate or that today’s systems have solved the credibility problem. They have not. It does mean the incentive to understand the entities behind the content becomes more significant. An AI assistant does not need another fifty companies claiming to be experts. It needs enough reliable information to determine which companies are relevant, capable and appropriate for the question being asked. Visibility and trust are related, but they are not the same thing.

When Everybody Can Claim Expertise, Brands Matter More

There is an apparent paradox here. If AI makes it easier to manufacture a professional identity, you might expect brands to become less important. I think the opposite is more likely. A brand is not merely a logo, a color palette or a clever tagline. Those are elements of presentation. A meaningful brand is the accumulated understanding of an entity: what people associate with its name, what customers expect from it, what work it is known for and what reputation it has built over time. When content becomes abundant, a recognizable identity helps people make sense of the abundance. When almost anyone can claim expertise, the history and reputation associated with a name can become more valuable, provided they are supported by something real.

This is one reason businesses should think beyond the narrow objective of getting a particular page to rank. Search visibility remains important, but a durable digital presence involves more than a collection of optimized URLs. A business needs a coherent identity, useful content, a well-organized website, evidence of its work and a reputation that is not entirely dependent on its own marketing. That is the larger relationship between SEO and building a credible business presence: technical optimization and content help search systems understand a company, but the company still needs substance worth understanding. A strong brand gives those signals a recognizable center. It makes the business easier to identify, evaluate and remember.

The real experts may even gain an advantage as the difference becomes clearer. People who have genuinely spent years doing something tend to accumulate material that cannot be created simply by asking an AI to invent a convincing professional history. They have solved problems, made decisions, learned from mistakes, developed opinions through experience and worked with people who can speak about the results. They have projects, relationships and a history of decisions with consequences. AI can help them explain those things, organize them and turn their experience into useful resources. What it cannot do is retroactively give a person twenty years of firsthand experience merely by generating a biography that claims they have it. History matters because it connects present-day claims to a sequence of real events.

The Answer Isn’t to Manufacture Credibility Faster

Businesses watching these changes may be tempted to respond by producing more content, publishing more frequently and creating an even more polished presentation. Sometimes those efforts are worthwhile. A neglected website needs attention, and a business with valuable experience should not hide it behind a poor digital presence. But trying to solve a credibility problem by manufacturing more signals of credibility is an arms race with no obvious finish line. If everyone publishes ten AI-assisted articles a week, publishing eleven is not necessarily a competitive advantage. If every competitor claims to be an expert, making the claim louder does not establish expertise. The objective should be to build something that can withstand inspection: show the work, explain the decisions, document real projects, discuss problems actually solved, publish original observations and make genuine customer experiences easier to discover.

This is where professional web design still has an important role. A website should not merely make a business look impressive; it should make the business understandable. It should present the work clearly, explain the services accurately, connect claims to supporting detail and give visitors a coherent picture of what the organization actually does. Good design does not need to disguise reality. It should make reality easier to evaluate. That distinction matters because a business with substantial experience and a terrible website may be harder to understand than it should be, while a business with no meaningful track record can use a beautiful website to create an impression that its experience does not justify. The opportunity is to bring presentation and substance together rather than treating one as a substitute for the other.

The Future Is Not Less AI. It Is More Evidence.

I don’t pretend to know exactly how the next generation of search algorithms will work. The systems are complicated, the signals change and the companies building them do not disclose every detail of their methods. But the broader pressure is easy to understand. As synthetic information increases, verification becomes more valuable. As generated content becomes abundant, firsthand experience becomes a more useful differentiator. As polished websites become easier to create, independent reputation can provide additional context. As more people claim expertise, evidence of that expertise becomes more important to anyone trying to decide whom to trust. These are not guarantees about a particular ranking factor or a prediction that search engines will adopt one specific test. They are the consequences of a digital environment in which the appearance of authority is becoming easier to produce at scale.

AI does not necessarily destroy trust. It raises the stakes of establishing it. The same tools that allow somebody to manufacture a convincing business identity can help a genuine business communicate its real experience far more effectively than it could before. A small company can explain its process, document its projects, produce useful educational material and make its expertise accessible without employing an entire marketing department. That is a substantial opportunity. But it works best when the technology amplifies something that exists rather than inventing something that doesn’t. The website is the presentation layer. Search is part of the discovery system. Reputation supplies context, and evidence helps connect the claims to reality. Those pieces work together; none can reliably substitute for all the others.

Eventually, we may reach a point where a perfectly polished website is no longer particularly impressive on its own. Not because good design has stopped mattering, but because it has become so accessible that it tells us less than it once did. A flawless brand identity, a beautifully written article and a professional-looking biography will establish that somebody knows how to present themselves. They will not necessarily establish that the person knows what they are doing. The natural follow-up question becomes increasingly important: Who are these people? What have they actually done? Who has worked with them? What exists because they were there? What happens when you look beyond the marketing? Those are the questions that begin to separate appearance from substance, not by judging which looks more impressive but by looking for the roots beneath the surface.

That is ultimately why I think the people actually doing the work have a real opportunity in this next phase of the web. AI can make a nobody look like a somebody, at least for a while. It can manufacture the language, the imagery and the professional presentation that once required a much larger investment. But the businesses doing genuine work have something worth bringing into view: projects, customers, lessons, decisions, experience and a history that connects their claims to the world outside their websites. They can use AI to communicate those things more effectively without pretending they have done something they haven’t. The goal is not to build a web without AI. It is to build a web in which generated presentation is no longer mistaken for proof.

The bullshit isn’t going away. If anything, it is going to get much, much better. The real question is whether search engines, AI systems and the rest of us get better at seeing through it.