Humans Are Better At Writing Than AI In These Tasks

Enhance your expertise with Progress Memo’s weekly knowledgeable insights. Subscribe without spending a dime!

There’s something ironic about making an attempt to make AI content material extra human. However there’s additionally one thing thrilling about it as a result of our work as writers and content material creators adjustments essentially.

This shift jogs my memory of my time as a DJ – many moons in the past.

I got here up within the period when DJs would haul eight to 10 crates of vinyl data to each gig. Throughout units, I’d frantically dig by these crates, looking for the right subsequent observe.

Like a author drawing from their psychological library of phrases and concepts, I needed to bear in mind the place particular data have been and develop my very own tagging system to seek out them sooner.

Then, Serato¹ modified every little thing.

This new expertise permits you to use two particular vinyl data to play your total digital music assortment. No extra hauling crates – any track was immediately accessible.

The sport modified fully.

Whereas some benefits disappeared (the exclusivity of getting uncommon data), new inventive potentialities emerged (like seamless remixing and creating customized edits).

The identical transformation is going on to writing right now: LLMs are our Serato.

As a substitute of laboriously crafting each sentence from scratch or drawing solely from our psychological archives, we are able to immediately entry numerous expressions and concepts.

Like digital DJing instruments, AI writing assistants give us an enormous inventive palette to work with.

And writers ought to embrace this shift! There’s no particular honor in doing issues the laborious approach.

Nonetheless, simply as a Serato DJ nonetheless wants musical data and efficiency expertise, uncooked AI output nonetheless wants human refinement.

With out cautious modifying, AI-generated content material feels sterile and impersonal, making it much less more likely to resonate with readers or carry out properly on social platforms.

The bottom line is treating AI-written materials as a place to begin – uncooked tracks to be combined, if you’ll – after which thoughtfully enhancing it to create one thing really compelling and human.

At its core, this piece about modifying AI content material is absolutely about one query: “What’s the distinctive worth people can nonetheless add to content material?”

Picture Credit score: Kevin Indig

After numerous nights wrestling with this query each intellectually and emotionally (hey, 3 a.m. nervousness!), I believe I’ve lastly cracked the code.

Primarily based on my expertise and from hours of “mixing” with AI, I’ve recognized seven uniquely human writing capabilities that no AI can genuinely replicate:

1. Patterns: Detecting refined alerts in wording, rhythm, and analogies that resonate with shared human expertise.
2. Matters: Intuitive understanding of what readers will discover genuinely fascinating or related.
3. Expertise: Private tales and views, particularly from people with established reputations.
4. Judgment: Making use of nuanced ethical reasoning past programmed tips.
5. Style: Making selections about what works stylistically.
6. Richness: Describing tastes, smells, textures, and sensations from lived expertise.
7. Secrets and techniques: Incorporating insights or information not obtainable in AI coaching units.

Whereas people typically overestimate true novelty (most progress is incremental), we stay the important curators of AI output.

Possibly what issues most isn’t whether or not AI created one thing however whether or not people acknowledge its worth.

Picture Credit score: Kevin Indig

To distinction the strengths of AI vs. people, I wish to break it down throughout 4 classes:

1. Technical execution.
2. Information & data.
3. Manufacturing & adaptation.
4. Knowledge & emotional intelligence.

1. Technical Execution

AI:

• Processing and error prevention (grammar, spelling, consistency).
• Sustaining strict formatting throughout lengthy paperwork.
• Following detailed content material tips with precision.
• Producing grammatically correct content material at scale.

People:

• Breaking established guidelines in significant, progressive methods.
• Creating new kinds, codecs, and genre-bending approaches.
• Growing distinctive private writing kinds.
• Writing with an genuine voice from lived expertise.

2. Information & Data

AI:

• Synthesizing data from huge data bases.
• Producing factual content material when based mostly on information.
• Creating complete explanations of advanced matters.
• Cross-referencing data from a number of domains.

People:

• Contributing authentic analysis and firsthand observations.
• Growing genuinely novel philosophical insights.
• Creating work pushed by genuine ethical conviction.
• Writing from a deep cultural understanding of particular communities.

3. Manufacturing & Adaptation

AI:

• Producing excessive volumes of content material shortly.
• Creating variations on current themes and codecs.
• Translating between languages with excessive accuracy.
• Restructuring content material for various audiences and platforms.

People:

• Inventing solely new literary kinds and approaches.
• Crafting narratives that reply to the cultural second.
• Creating humor that depends on nuanced cultural context.
• Growing satire that addresses modern points.

4. Knowledge & Emotional Intelligence

AI:

• Changing structured information into readable narratives.
• Summarizing prolonged content material whereas preserving key data.
• Creating constant documentation from technical specs.
• Adapting content material throughout a number of codecs and channels.

People:

• Creating characters with advanced, contradictory motivations.
• Writing dialogue that captures psychological nuance.
• Conveying refined emotional states by deliberate phrase alternative.
• Crafting tales that evoke highly effective emotional responses.

People ought to let AI deal with the baseline 80% – the beat-matching and tempo management, if you’ll.

And we should always focus our inventive power on that vital 20% the place we combine within the samples that no person else has:our distinctive views, stunning tales, ethical nuance, cultural references, and really novel concepts.

Now that we perceive what makes human content material useful, we have to acknowledge what makes AI content material really feel off-putting.

Simply as an newbie DJ may technically match beats however nonetheless create an ungainly, lifeless set, AI writing has particular patterns that sign “one thing’s not fairly proper right here.”

Bear in mind how even digital DJs nonetheless must learn the room?

Equally, whereas fashionable language fashions can technically string phrases collectively fantastically, they nonetheless miss essential human alerts.

Via deep analysis and shopper work, I’ve recognized 11 telltale indicators that scream “an AI made this” – patterns that immediately disconnect readers:

• Sterile language: Overly formal phrasing that no human would truly use.
• Structural monotony: Predictable sentence patterns that create a hypnotic rhythm.
• Awkward transitions: Abrupt jumps between concepts with out pure connective tissue.
• Robotic tone: An impersonal voice that retains readers at arm’s size.
• Factual shakiness: Assertions that sound believable however don’t maintain as much as scrutiny.
• Character vacuum: Writing devoid of quirks, humor, or genuine perspective.
• Generic protection: Floor-level remedy of predictable matters.
• Sourceless claims: Knowledge statements with out correct attribution.
• Shallow insights: Concepts that by no means push past the apparent.
• Model misalignment: Content material that doesn’t match your established voice.
• Weak bookends: Forgettable openings and conclusions that fail to have interaction.

The markers actually stand out.

Within the paper “Linguistic Markers of Inherently False AI Communication and Deliberately False Human Communication,”²researchers have been capable of detect 80% of AI content material precisely by on the lookout for:

• Extra emotional/affective language.
• Extra analytic writing type.
• Extra descriptive (greater use of adjectives).
• Much less readable (extra advanced sentence buildings).

Paradoxically, many human writers show these identical weaknesses.

The distinction? People can study to beat them.

The DJ analogy actually comes full circle right here.

Simply as the most effective DJs don’t merely play songs in sequence however create one thing new by their mixing, the best content material creators don’t simply edit AI output – they remodel it.

In right now’s panorama, probably the most useful content material comes from creators who:

1. Perceive the place AI instruments excel (the technical baseline).
2. Acknowledge the place human enter is important (these seven distinctive capabilities).
3. Can establish and remove these telltale AI patterns.
4. Know methods to mix the 2 seamlessly into one thing better than the sum of its components.

We’re not simply modifying AI content material – we’re remixing it with our uniquely human perspective, creating one thing that no algorithm may generate alone.

As a result of in the end, probably the most compelling content material doesn’t come from people combating in opposition to AI or from AI trying to exchange people. It comes from the considerate collaboration between each.

Subsequent week, I’ll break down my precise workflow for modifying AI content material – the sensible methods I exploit every day to remodel sterile AI output into content material that genuinely resonates, connects, and performs.

You’ll learn to effectively leverage these instruments whereas making certain your content material maintains that irreplaceable human contact.

¹ Serato

² Linguistic Markers of Inherently False AI Communication and Deliberately False Human Communication: Proof From Resort Opinions

Featured Picture: Paulo Bobita/Search Engine Journal