How I make UNDETECTABLE AI Content (that Google Loves)
How I make UNDETECTABLE AI Content (that Google Loves)
Are you wondering how to create undetectable AI content that Google will love? In this video, you’ll learn how to produce high-quality, undetectable AI articles that will become a massive traffic generator.
Timestamps:
00:00 Introduction
00:48 Generate AI Content
03:43 Free Techniques
05:10 Write as Human
07:33 Emoji \u0026 Comma Tricks
09:09 Why This All Matters
Resources:
Open.ai Text Classifier: https://platform.openai.com/ai-text-c…
Originality.ai: https://originality.ai/
The Affiliate Lab: https://affiliatelab.im/
AI Content Website: • I Built a Website with 100% AI Conten…
Surfer AI Beta Test: https://forms.gle/LkFxWX9tviLT8Et47
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Subscribe here
►► https://www.youtube.com/MattDiggity?s…
Download my Evergreen Onsite SEO Guide FREE
►► https://diggitymarketing.com/onsite-s…
For more Advanced SEO Tips, check out this Playlist
►► • Advanced SEO Tips, Tricks, and Techni…
Content
0 -> in this video I'm going to show you how
1.5 -> to create completely undetectable AI
3.959 -> content Google will have no clue if it's
5.759 -> written by an AI and neither will humans
7.5 -> want proof an AI content piece pulled a
10.019 -> 100 guaranteed AI score before I worked
12.66 -> my magic on it then after I use just one
14.58 -> of the techniques I'll show you in this
15.839 -> video we now have 100 original Okay cool
18.359 -> so I can pass AI detection tools so what
20.82 -> how does that translate into anything
22.199 -> useful well when you can produce high
24.119 -> quality undetectable AI articles you can
26.699 -> generate enormous amounts of content
28.32 -> like it's no one's business which is
29.82 -> going to sprout a massive traffic boner
31.5 -> for your website so you can then start
32.94 -> printing money like Papa Powell now
34.68 -> there's multiple ways that people claim
36.36 -> they can beat AI detection tools such as
38.399 -> rewriting the content with quill bot the
40.2 -> Emoji trick and quite a few more I'm
42 -> going to test each of them out to see
43.62 -> which ones work but first let's generate
45.78 -> some AI content and plug it into each of
47.82 -> the detection tools so we can get a
49.26 -> baseline to avoid chat GPT lag time and
51.719 -> usage blackouts which have been rampant
53.399 -> as of late I'm going to use open ai's
55.199 -> playground to generate my test content
56.76 -> here's my prompt and thank me later if
58.559 -> you're an 80s fan write me an article on
60.6 -> my Big Trouble in Little China is a
62.28 -> better movie than Top Gun bam here's the
64.14 -> content now let's go on over to
65.54 -> originality.ai to see if it Buffs me
67.619 -> red-handed for using AI this is a paid
69.6 -> tool so let's hope it diagnoses it
71.28 -> correctly you plop in the content here
72.9 -> let it run and here we go originality
75.06 -> says there's a 98 chance of being AI
77.4 -> content good for you originality now
79.2 -> let's toss it into AI cheat sheet.com
81.119 -> one of the free tools as expected 100
83.28 -> certain the text was written by Ai and
85.619 -> lastly let's test out open ai's own
87.36 -> detection tool AI text classifier okay
89.82 -> then the classifier considers the text
91.619 -> to be very unlikely AI generated pretty
93.96 -> fast that open ai's own tool misses the
95.88 -> mark But noted so here's where we stand
97.799 -> Article 1 is 98 likely to be AI on
100.92 -> originality 100 AI on cheat sheet and
103.799 -> unlikely an open ai's own tool let's
105.72 -> keep going and create a few more test
107.1 -> cases keeping up with the Retro theme
108.84 -> open AI write me an article on why
110.64 -> rollerblading is dead four percent
112.259 -> chance it was AI according to
113.7 -> originality a hundred percent on AI
115.74 -> cheat sheet and unclear according to
117.42 -> open AI so putting all these results
119.04 -> together so far we're starting to see
120.78 -> some inconsistency at least with the
122.399 -> first tool open a is tool is still lost
124.2 -> AF before you start tricking these tools
126.18 -> let's generate one more test case write
127.979 -> me an article on proper form for bench
130.02 -> press AI D even lift bro apparently it
132.66 -> does and here's the content so this time
134.52 -> we have originality with 99 Ai and cheat
137.4 -> sheet with one percent and open AI
139.14 -> itself still confused here's our
140.76 -> Baseline ultimately both originality and
142.92 -> AI cheat sheet are inconsistent and open
144.72 -> ai's tool can probably just be ignored
146.459 -> let's start to do some magic tricks to
147.959 -> see if I can improve these scores but
149.879 -> first make sure to sign up for my free
151.56 -> SEO training masterclass by using the
153.36 -> link in the pin comment it goes over all
155.28 -> the magic I'm doing today to get sites
157.02 -> to the top of Google Now back to the
158.52 -> show the first AI undetection technique
160.68 -> is that even a word anyways the first
162.48 -> technique I'm going to use is quill bot
164.04 -> for rewriting your content you do do
165.959 -> this by going to quill bot dropping in
167.64 -> your content and then hitting paraphrase
169.5 -> once I put in the 80s movie AI text we
172.019 -> get this new content on the right let's
173.64 -> just copy this new text and paste it
175.379 -> back into originality well well well
177.12 -> originally said it was 98 fake before
179.64 -> now it's 95 real that was easy now let's
182.819 -> drop into the AI cheat sheet 100 AI
185.28 -> content before and now it's 59 chance of
187.92 -> its human written or in other words 41
189.9 -> chance of his AI then we have open AI
192.42 -> itself with possibly AI generated that's
194.58 -> actually an improvement for the tool
196.14 -> from their previous unlikely
197.519 -> classification bringing the same quill
199.44 -> bot process over for the two other test
201.12 -> cases we end up with the following
202.739 -> results which are interesting if the
204.48 -> Articles were previously super
205.86 -> detectable by the tools then coolbot is
207.959 -> for sure going to improve it but if they
209.64 -> were already passing with flying colors
211.14 -> cobot will likely make it worse which is
213.18 -> expected if it's not broken don't try to
215.04 -> fix it coolbot requires a paid account
216.78 -> if you input any significant amount of
218.7 -> characters so let's move over to testing
220.44 -> some of the free techniques the
221.879 -> following few are all under the umbrella
223.68 -> of telling the AI not to write like an
225.599 -> AI starting with perplexity and
228 -> burstiness what
230.28 -> there's this post on medium where the
232.08 -> author shared the following prompt to
233.76 -> pre-train your AI before it starts word
235.68 -> vomiting hey chatgpt regarding
237.78 -> generating writing content two factors
239.76 -> are crucial to being the highest degree
241.5 -> perplexity and burstiness perplexity
243.84 -> measures the complexity of the text
245.58 -> separately burstiness compares the
247.739 -> variations of sentences humans send them
249.9 -> right with greater burstiness for
251.459 -> example with some longer or more complex
253.26 -> sentences alongside shorter ones AI
255.72 -> sentences tend to be more uniform
257.28 -> therefore generated text content must
259.44 -> have the highest degree of perplexity
261.06 -> and the highest degree of burstiness the
263.28 -> other two factors are that writing
264.66 -> should be maximum contextually relevant
266.759 -> and maximum coherent take this word
268.68 -> block drop it into your AI tool and then
270.78 -> ask it to create a Content based on the
272.82 -> above write me an article on why Big
274.56 -> Trouble in Little China is a better
275.94 -> movie than Top Gun originally says it's
278.1 -> 78 AI which is an improvement from the
280.32 -> 98 before AI cheat sheet went from 100
282.96 -> AI to only 11 Ai and open AI itself went
286.68 -> from possibly to likely adding in the
288.72 -> results for the other two test cases we
290.1 -> end end up with the table like this I'd
292.02 -> say that this perplexity prompt is
293.699 -> pretty damn inconsistent it did decently
295.56 -> for the first test case but bombed ass
297.24 -> on the second two Skip by the way if
299.16 -> you're curious as to why we even care in
300.78 -> the first place about passing a
302.04 -> detection software first off yes it's
304.44 -> very important and I'll dig into why
305.94 -> after I finish up these tests next we
307.86 -> have another pre-training prompt that
309.479 -> tells the tool how to write like a human
311.1 -> try to sound like a human writer writing
313.199 -> for a Blog writing in the first person
314.82 -> giving advice try to sound unique and
317.04 -> right in an unpredictable fashion that
318.9 -> doesn't sound like gpd3 here's how I
320.94 -> perform for test cases one and two it
322.62 -> screwed the pooch but for the bench
324.06 -> press content it did really well how
325.8 -> about if we ask the AI itself how to
327.72 -> write undetectable AI content what are
329.82 -> the key attributes of conversational
331.199 -> content that is undetectable as being
333.3 -> written by an AI open AI tells you that
335.58 -> this content would include human-like
337.38 -> grammar punctuation and spelling that
339.12 -> would look human contact space content
340.979 -> creation and so forth and then you ask
342.96 -> it to generate your content now we're
344.759 -> talking based on all the simple
346.02 -> pre-training exercises we've used so far
348.06 -> asking the ad directly how to pass the
349.979 -> tech gives the most consistent result
352.02 -> what if we ask the almighty AI how to
354.12 -> fix content that's already written let
355.8 -> it first write your content in vanilla
357.539 -> mode then you ask it to rewrite the
359.1 -> above content so that it is not detected
361.259 -> as AI content by a Content writers and
363.66 -> here's a result winner winner chicken
365.22 -> dinner this is our best performing
366.9 -> technique so far simply asking the AI to
369.36 -> rewrite itself for passability results
371.1 -> in great scores for all the tools
373.979 -> foreign
375.9 -> thank you to affiliate lab member and
377.759 -> Chiang my buddy chrismanac for that tip
379.56 -> next if you haven't already seen my case
381.18 -> study where I broke 50k traffic with a
383.28 -> pure AI site make sure to check it out
385.08 -> after you finish here Link in the
386.4 -> description in that video I mentioned
388.02 -> that I've been using a beta version of
389.639 -> surfer's upcoming AI tool that will
391.68 -> generate AI content on the Fly that's
393.6 -> actually SEO optimized it looks at the
395.699 -> top ranking articles to determine how to
397.5 -> create the article outline how long to
399.18 -> write the content and what entities and
400.979 -> critical keywords need to be in your
402.479 -> content and at what frequencies it
404.46 -> actually does a lot more than that but
405.66 -> let's see how detectable it is bear in
407.4 -> mind the surfer tool generates full
408.96 -> articles and longer content has a much
411 -> higher chance of detectability so here
412.86 -> goes nothing the big trouble article hit
414.84 -> 28 AI on originality AA cheat sheet
417.96 -> guarantees it was a human who wrote it
419.699 -> in the open AI tool choked up on it
421.56 -> probably because it's too long that's
423.06 -> what she said
425.639 -> ladies and gentlemen woman we have a new
427.139 -> Champion the surfer content bamboozle
429.06 -> seat detection tool is better than any
430.68 -> other technique so far now there's
432.06 -> something I want to mention I gotta
433.5 -> admit that was more than a bit nervous
435 -> putting the surfer tool to the test not
436.86 -> only because I'm an investor but because
438.36 -> I'm putting this content on my freaking
439.979 -> websites so seeing these results was a
441.9 -> huge relief and if anyone would like to
443.52 -> verify that I'm not making these numbers
444.96 -> up I'm happy to share the content with
446.699 -> you so you can run the tools yourself if
448.319 -> you're interested in applying for the
449.639 -> surfer AI beta test use the link in the
451.5 -> description the next Technique we have
453.06 -> is a fun one and that's the Emoji trick
454.979 -> you ask it to generate your content but
456.84 -> insert an emoji after each word and then
459.06 -> when it's done you ask it to remove the
460.56 -> Emojis while the Emoji trick did perform
462.419 -> Super well in the bench press article it
464.4 -> definitely fell short in the other two
465.599 -> in my opinion this isn't consistent
467.4 -> enough so let's call this mythbusted
469.02 -> next the comma trick write me an article
470.94 -> on proper form for bench press but
472.8 -> remove as many comments as possible the
474.78 -> combat techniques seem to work on
476.099 -> originality and open ai's detector in a
478.199 -> test case or two but it's definitely not
479.759 -> consistent enough to add to your process
481.56 -> now we're about to get into why this all
483.539 -> matters but let's take a look at the
485.099 -> final results the winners are clear
486.72 -> Surfers AI performs the best followed by
489 -> the rewrite the content technique and
490.8 -> then good old you tell me how to write
492.479 -> human-like content Google has gone back
494.52 -> and forth on their stance on whether or
496.139 -> not AI content is against their
497.699 -> guidelines or not in April of last year
499.44 -> they were like nah bro AI is fully
501.479 -> against our guidelines specifically if
503.34 -> you're using machine learning tools to
504.9 -> generate your content it's essentially
506.46 -> the same as if you're just shuffling
507.9 -> words around or looking up synonyms or
510.24 -> doing the translation tricks that people
511.74 -> used to do and that doing that is still
513.659 -> automatically generated content and that
515.339 -> means for us it's still against the
516.899 -> webmaster guidelines but then in
518.52 -> February they released an official
520.02 -> statement in Google search Central and
521.7 -> said they'd reward high quality content
523.62 -> however it's produced okay but that's
526.62 -> not surprising they have to take the
528.36 -> stance because they're clearly going to
529.8 -> be used AI themselves in their search
531.36 -> results to compete with Bing's chat GPT
533.519 -> integration and Danny Sullivan clarified
535.68 -> this with quote content written
537.6 -> primarily for search engines rather than
539.58 -> humans is the issue if someone fires up
541.8 -> a hundred humans to write content just
543.6 -> to rank or fires up a spinner or AI say
546.24 -> issue sure I agree that the ultimate
548.94 -> goal would be to create content that
550.8 -> humans would enjoy and encourages them
552.839 -> to do what you would want them to do on
554.459 -> your website namely generate you money
556.2 -> through ads or conversions But
557.82 -> ultimately as a search engine
559.14 -> professional I'm obviously producing
561.24 -> this content so it performs well in
562.86 -> Google or any other search engine for
564.6 -> that matter but there seems to be an
566.22 -> unclear line that you're not supposed to
567.839 -> cross in the affiliate lab we have a
569.76 -> private Mastermind group with the top
571.32 -> performers who have sold businesses for
573.06 -> six figures or more one of the members
574.8 -> Dave Gibbons has a theory which I
576.42 -> gravitate to the most Google will treat
578.22 -> it just like it does with plagiarism or
580.26 -> duplicate content AI will be allowed but
582.42 -> it will be deemed a lower quality
583.74 -> article than a Unique Piece with
585.42 -> equivalent other SEO factors I agree 100
588.18 -> Google needs to incentivize people for
590.519 -> creating real original content otherwise
592.8 -> what in God's name will the AI be able
594.959 -> to scrape its information from so if
596.7 -> you're going to be using AI in your SEO
598.44 -> strategy the name of the game is to make
600.54 -> it appear human I told Dave that this
602.459 -> was a very good theory after which he
604.14 -> sarcastically responded with if only I
606.12 -> was in a group with someone with an
607.44 -> engineering background that could build
608.76 -> a hypothesis from it and do a load of
611.04 -> tests and then let me know the reality
612.42 -> touche my friend touche and that's what
614.76 -> I'm up to right now I'm running single
616.32 -> variable tests right now to see how AI
618.12 -> content performs against human content
620.16 -> in the actual search results make sure
622.26 -> to subscribe so you don't miss my
623.58 -> findings
Source: https://www.youtube.com/watch?v=Ryr15qNIdrw