Conference shows that the best and worst aspects of the internet are separated only by missing conditions

The assertion of an 'internet death' is repeated quite often, although previously, in August 2010, Wired published a cover declaring the demise of the Net. As noted in The History of the Web, critics, including Rob Besiczu, pointed out that the graph used for this statement only showed an increase in the internet's share of the total volume, not an actual decline in usage.

However, this time, according to Wired, those who mourn may have grounds, and the problem goes beyond mere traffic. Wired returns to the topic, presenting new material based on the book by journalist and technologist John Vest titled The Internet Will Die, and So Will You. The main complaint is that the positive aspects of the Net no longer outweigh the amount of useless content.

The problem lies not in the infrastructure, as the cables function normally, but in the user experience itself. One expert notes: 'Any rational observer can admit that the current version of the Net—and the social structure woven within it—does not match what people initially expected.'

When searching for a simple cooking instruction, a user might receive a four-thousand-word memoir about someone's grandmother, three automatic videos, and a banner with numerous options. When browsing a feed, the user is offered content designed to hold attention, not the content they were looking for. Cory Doctorow has named this slow decay 'enshittification,' which is an accurate description of the situation.

Furthermore, there is the issue of 'slop.' Thanks to artificial intelligence, machine-generated content has become so cheap that content farms can publish materials faster than a person can type 'I'm not interested.' There is also a fringe theory of a dead internet, suggesting the Net primarily consists of automated accounts communicating with each other. Although the strong version of this theory is considered conspiratorial, evidence supporting a milder version is accumulating.

Wired references a cybersecurity firm's study showing that AI agents now exceed the number of human users, as their traffic has increased by approximately eightyfold by 2025. Bots have stopped being mere guests and have begun taking control of the process.

Another aspect is the business model. Wired notes that Oxford University Press named 'rage bait' word of the year in December after its usage increased by about threefold. This indicates that outrage generates profit, and AI allows scammers to produce it in huge quantities.

John Vest's arguments become particularly interesting. He believes the Net is far more fragile than it appears. His book begins by describing a 2001 railway accident in the Baltimore tunnel that damaged the cables transmitting internet traffic and slowed connections across the US. The problem of 'link rot' is also present.

Wired cites two examples from Vest: approximately one in four links in New York Times articles from 1996 to mid-2019 is no longer working, and a Pew report found that over half of Wikipedia entries contain at least one broken link. Thus, humanity's permanent archive is becoming a strategy of decaying content.

Vest gives a personal example: his teenage anime fan page is still active and filled with spam because he lost access to the account needed to delete it. His university blog lost a year due to an unpaid domain account. Vest's conclusion, according to Wired, is that the concept of permanence is a myth, and useless content tends to outlive monuments.

The reason this should be allowed to happen is that the bloated, ad-saturated, algorithm-driven version of the Net is dying, and no one should try to resuscitate it. Instead, attention should be paid to what is growing in the 'cracks'—newsletters landing in inboxes without algorithmic approval; forums managed by passionate people genuinely interested, for example, in vintage fishing videos; group chats that serve as the last functioning town square in South Africa and have better moderation than most multi-billion dollar platforms.

According to Vest, content also thrives in unnoticed corners: obscure YouTube videos watched by almost no one, and web poetry journals distributed out of love rather than for money. The pattern is simple: meaning is best preserved where there is no financial gain.

A small 'dip' for all this junk won't hurt. It is worth turning off content factories for a few hours and seeing who notices. Most likely, the lights will remain on in the places that truly matter.

In conclusion, for publishers, the decision is to stop writing for the algorithm and start writing for the person on the other side of that algorithm. One must have a unique voice, cover real events, and write in a browser with a specific purpose. If the entire strategy boils down to being one of the links in search results, the inevitable funeral of the internet is guaranteed, and you will be on the guest list.

Modern daily interactions—whether it's a message from a family member, an email from a manager, or a friend's advice after a quarrel—increasingly involve artificial intelligence. This conversational participant does not experience emotions nor has personal vested interest in the relationship.

For many generations, people learned social etiquette, understood different viewpoints, and navigated relationships through real interaction, learning from mistakes and misunderstood intentions.

However, now that AI can instantly provide answers to any social dilemma, some of these efforts can be delegated to a machine. The question arises: why do ordinary messages require assistance, and does AI make us better at forming connections, even if we become more eloquent, polite, and confident?

Humans are traditionally considered social beings, and relationships are shaped not only by what is said but also by how people listen, react, disagree, and adapt to each other. AI offers something rarely achieved in human relationships: immediate attention, infinite patience, and responses tailored to specific issues. It lacks awkward silences, raised eyebrows, or the fear of embarrassment.

Shailesh Nevatia, founder and director of Grandeavour Communication, who spent nearly 25 years in the communications field but described himself as shy in public speaking, noted this characteristic. He stated that AI helps him handle disagreements by offering perspectives he might otherwise miss. He emphasized: 'With people, we may have certain opinions, views, or thoughts, whereas the most valuable thing AI offers is not the right answer, but a different perspective on things.'

Nevertheless, he insisted that the emotional depth, vulnerability, and shared experience of human relationships cannot be replicated by a chatbot.

Concerns about communicating with AI are related not only to people using it to find the right words but also to the fact that constant dependence on it may undermine their own confidence in finding those words.

For instance, Riddhi almost instinctively turns to ChatGPT when drafting replies. Whether she needs to sound politely firm in a business email or find a way to decline plans with friends, AI assists her in this situation. She admitted: 'It's not that I can't reply without AI, but having someone to check the correspondence is better than just hearing how it sounds in my head.'

However, this confidence comes at a cost. Riddhi added: 'AI understands my writing so well that I'm not very sure even when writing this, which is one of the biggest drawbacks.'

Despite this, she called for embracing AI, but doing so cautiously. Since AI has become the new normal, returning to an era without it is impossible, but it is important not to let it become its master.

Riddhi's experience revealed a potential paradox: a tool designed to simplify communication might reduce users' confidence in their ability to communicate independently.

Risks increase when AI is used not just to formulate a response, but to interpret another person's intentions. Consider a situation where a person consults a chatbot after a fight to determine who is at fault. The answer may sound thoughtful and convincing, but it is based on the version presented to the system, which may overlook details or reflect the user's own biases.

In this context, Resmita Nair, a counseling psychologist at SPARSH Hospital, warned that such interactions can reinforce confirmation bias, causing people to believe their existing interpretation more strongly rather than being open to another point of view.

This is important because disagreements are also opportunities for developing social skills. Understanding another person's perspective, managing unpleasant emotions, and responding to unexpected reactions require practice, not just a well-formulated answer.

Nair explained: 'Social skills are acquired in real situations that involve emotions, misunderstandings, conflict, and sometimes rejection.'

Constantly delegating such situations to AI can diminish opportunities to develop empathy, emotional intelligence, and the ability to cope with uncertainty. A person might learn to phrase a difficult message more effectively, but this doesn't necessarily make them better at the subsequent conversation.

A message is not just a set of sentences. Its meaning can change depending on the relationship between people, their shared experiences, and even the pauses between words.

Ritu Mehrotra, founder and CEO of Shunya Labs, a voice AI company, noted that natural communication involves much more than just language. She said: 'Communication is not just words; it is also tone, pauses, context, emotions, and all those things we convey without explicitly saying them.'

Progress in voice AI allows technologies to interpret language, intent, and context with greater sophistication, including different accents and communication styles. However, recognizing emotional signals is not the same as experiencing emotions. Mehrotra stressed: 'As AI becomes part of more and more personal conversations, we must realize that understanding emotions is not experiencing them. Trust, privacy, and transparency must be fundamental.'

AI should help people communicate with each other, not imitate human empathy or replace human relationships.

A chatbot can recognize expressions of sadness or affection and generate a response that sounds sympathetic. But recognizing emotional patterns is different from sharing emotional responsibility in human relationships. For example, when a chatbot was asked for advice on responding to a hypothetical message to a friend who had been laid off, the response was courteous but generalized. It offered comfort without knowing the friend's emotional state, personality, or the nature of the relationship.

Even a seemingly caring question like 'Are you okay?' can sound inappropriate if the person is seeking support rather than an answer to an obvious question. A more personal response might acknowledge the setback while affirming the friend's abilities: 'Oh god, that's terrible. I know the job market is tough, but you are talented. You will find your place.'

The difference is not simply choosing warmer words. It is in understanding what a specific person might need at that moment, which a chatbot may fail to assess without sufficient context.

There is also the issue of authenticity. When a message is meticulously constructed by AI, it can effectively convey the intended information. But does it still sound like a human sending it? In an example where AI removes the human touch, another chatbot was asked: 'Someone close to me is grieving. I don't know what to say and I'm afraid of making it worse. What should I say?' It gave an answer that wasn't wrong, but was too robotic: 'I am here for you. I am sorry for your suffering. Speak if you want; silence is acceptable.'

These are perfect words, but they are not spoken by a comforting human due to their 'stiff tone.' Subtle hesitations, imperfect expressions, and familiar phrases that characterize human communication can disappear under polished language. The message may become grammatically more correct, but less personal.

Of course, AI cannot be dismissed as a threat to social skills, as it can help people who find it difficult to express themselves. For Pooja, a college student who is shy about speaking in front of the whole class or constantly asking friends to explain things, AI provides a space where she can ask questions without worrying about the reaction of others.

She uses it to understand complex topics and brainstorm, and sometimes asks ChatGPT to explain something as if she were ten years old. She notes: 'What I like most is that I can ask the silliest questions without feeling embarrassed. But I also feel that it's very easy to become dependent on it.'

However, she acknowledges the possibility of excessive dependence. Pooja says: 'AI can give you answers in seconds, but actual understanding and learning still depend on us.'

For others, AI serves as a sounding board before difficult conversations. It also helps people realize how their words can affect others. It can suggest more inclusive phrasing, point out potentially offensive expressions, and help users consider perspectives they might have missed. For those unsure how to communicate in diverse settings, this can be a starting point for more thoughtful interaction.

But knowing which words are appropriate is not necessarily the same as understanding why they are important. True inclusivity requires empathy and a willingness to understand another's experience, qualities that must go beyond the words suggested by a chatbot.

Furthermore, in such cases, AI may not provide the perfect answer, but it can help think in the right direction and suggest options the user might not have considered. In another hypothetical test, the chatbot was asked how to support a family member struggling with mental health without pushing them away. Although its answer was not entirely correct, it offered various approaches to consider, giving the person a starting point for reflection on how to respond.

Thus, the value of AI may lie not in providing the perfect answer, but in helping a person see possibilities they hadn't previously considered. The crucial distinction is whether people use these suggestions to develop their own judgment or simply follow them without question.

As Shruti Rajan Kappil, a lead quality specialist at the mental health platform Amaha, explained, the need to be heard and supported is fundamental to human relationships, but not every problem feels safe to discuss. Shame, embarrassment, and fear of judgment often prevent people from opening up, especially where stigma persists.

AI offers an alternative, but Kappil pointed out an important difference. 'Good support often requires honest disagreement or a gentle challenge from another perspective. AI tends to agree most of the time, and agreement is easily mistaken for help.'

She also noted that constantly seeking AI approval before sending a message can become a form of seeking comfort. While this brings immediate relief, constantly checking if your words are 'right' can gradually undermine self-confidence in judgment.

Shrishti Choudhary, Senior Manager of Regulatory Intelligence and Market Research at Bureau Veritas, uses AI daily to summarize regulatory updates. Increasingly, she also consults it on interpersonal communication matters. From defining emotional nuances to adding warmth and politeness to messages for colleagues and professional contacts—AI has become a useful companion. While AI can refine written communication, help anticipate reactions, and interpret conversational nuances, it cannot replace the soft skills, verbal cues, and life experience that shape genuine human connections.