AI Content — Finding Its Place in an Age of Abundance

Every new invention and tool ultimately finds its fit. AI sometimes hallucinates and produces low-quality output — but can you imagine life without it now? Its reception varies across social platforms, because people visit different platforms for different reasons and with different expectations.

On platforms like TikTok and Instagram — continuous feeds, short attention spans, fast scrolling — high-volume or AI-assisted content can work well, as long as it is engaging enough to stop the scroll.

But on intent-driven platforms, where users arrive with a purpose and spend more time, AI-generated content is less welcome — either rejected by users themselves, or discouraged by the platform’s own business model. In July 2025, YouTube tightened monetization for creators producing mass-produced, repetitive, low-originality AI content. The reason is straightforward: advertisers paying YouTube ultimately want results — clicks, conversions, actions. AI-driven content can flood feeds with shallow, repetitive videos, leading to boring or misleading experiences. Viewers disengage, spend less time on the platform, and that directly reduces ad opportunities. Brand safety is another factor — most companies simply don’t want their ads appearing next to low-quality or questionable content.

The same pattern plays out across Google’s broader ecosystem. AI has made it easy to churn out content for SEO and search rankings, but much of it is shallow and doesn’t build trust. Google Search increasingly looks beyond keywords and backlinks, paying closer attention to real user behavior — whether people stay, engage, and take action. When content fails that test, rankings drop. And even paid Google Ads campaigns become less effective and more expensive, because clicks don’t convert. This creates real economic pressure: to get results and keep costs under control, advertisers are increasingly pushed toward authentic, useful content that genuinely earns user trust — not content that merely exists.

As audiences are exposed to increasingly similar AI-generated outputs, patterns become predictable — the same phrasing, structures, and ideas repeated across articles and videos. This reduces perceived originality and makes it harder for any single piece to stand out. Over time, users become more selective, gravitating toward sources that demonstrate clear expertise, lived experience, or a distinct point of view — qualities that are harder to mass-produce.

Other areas where AI-generated content tends to fall short:

Emotional advertising — AI struggles to create genuine human connection. For example, a major brand’s AI-assisted holiday campaign faced criticism for lacking the warmth and storytelling depth audiences expected.

Customer communication — AI-generated emails and support responses can feel robotic and impersonal. When someone reaches out with a real problem, an automated reply that misses the emotional context often makes things worse, not better.

Corporate messaging & PR — Sensitive announcements, apologies, or press releases drafted by AI can easily come across as hollow or insincere — exactly the wrong note when trust is already on the line.

Comedy and satire — Humour is deeply cultural, contextual, and timing-dependent. AI can replicate the structure of a joke but routinely misses the instinct behind it — what to exaggerate, what to leave unsaid, and where the edge of acceptable really is.

Long-form narrative storytelling — Novels, screenplays, and serialised content require sustained character voice, thematic consistency, and emotional payoff across hundreds of pages or episodes. AI-generated long-form content tends to drift — characters flatten, subplots lose coherence, and the emotional arc weakens over length.

Music with lyrical depth — AI can generate technically competent lyrics and melodies, but songs that resonate culturally — that capture a moment, a movement, or a feeling — require lived experience behind the words. The difference between a song that plays and a song that stays is still very human.

Legal narratives and oral arguments — While AI handles document drafting reasonably well, constructing a persuasive narrative for a jury or framing an oral argument before a judge requires reading the room, adapting in real time, and understanding the human stakes involved — not just the legal logic.

Therapeutic and counselling content — Mental health conversations require genuine empathy, the ability to sit with silence, and judgment about when to push and when to hold back. AI-generated therapeutic content risks giving confident-sounding but contextually wrong responses at deeply vulnerable moments.

Cultural and indigenous storytelling — Stories rooted in specific cultural memory, oral tradition, or community identity cannot be authentically generated by a model trained on generalised data. The risk here is not just poor quality — it is misrepresentation.

Political speechwriting at its best — Functional political communication AI can manage. But the speeches that move people, that land at the right historical moment with the right moral weight — those require a writer who understands what is at stake beyond the words themselves.

AI can draft. Only humans can mean it.

Note: Edited with help from Claude.