Pangram Reviews 15

This company reviews substantial portions of my human-written academic dissertation as 100% AI generated- including my themes developed from 30+ interviews, my research questions generated 2+ years ago and highly specific to my study, and even more. I'm blown away at how off this thing is, but yet it's increasingly being used as the "go to" tool."

This dumbass site is genuinely just wrong. Multiple of my professors have used this site and told me the chances of it being wrong are 1 in 10,000. Well, I guess I should buy a couple of lottery tickets then because WOW. I have had to dumb down countless of my essays because this site thinks that any big word is AI. It's so stupid, and it is genuinely making humans dumber; I can't even write a good essay anymore without worrying about it being flagged as AI. If you see that your professor, boss, or whatever uses this site to check your work, pick a god and start praying, because only she knows how much time of your life you will have to waste trying to revise your writing for NO reason. 1/10.

I am being forced to use this AI detected tool for courses at Capella University. This tool give false positive results. I have screenshots of proof showing that when I put a single paragraph into the tool that was completely written by me without any AI help, I get the result of 100% human written. Now if I put all the paragraphs together into this tool, the result is 100% AI written. How can the same individual paragraphs submitted go from 100% human written to all paragraphs submitted together result in 100% AI written, with ANY CHANGES. This proves that the made up algorithms this program chooses to use, are inconsistent, inaccurate and just plain wrong. This program needs to be removed from being used a guide for College work.

UPDATE: After being given a second chance and phone calls with my Academic Coach, I have finally dumb down, literally DUMB DOWN, my paper enough to middle school terminology in order to get 100% humanized. Grammarly and Turnitin will flag your paper as 100% AI usage, so I have been advised not to use them. This was confirmed by my professor and academic coach.

I am under review with my GRADUATE school for using AI because of this program. I decided to go ahead and by a subscription to see how actually accurate this AI tool that is checking for AI actually is.
I have uploaded 18 documents ranging from personal essays to research proposals. EVERY single personal essay about MY life has come back as 100% AI (which is 10/18 of the uploaded documents), 4 out of 18 are research proposals and came back as Human-written. 4 out of 18 were random documents, such as MY college portfolio, a fundraiser message, an obituary for my cousin, and a short essay about my summer, ALL of which came back as 100% AI.
Basically, 14 out of the 18 documents I have uploaded have all been deemed as AI for the phrasing I am using and the potential similarity in sentences.
This program is bogus, I wouldn't recommend it for any college to use, any student to use, or any workplace environment to use. Everything that I have heard about this program and the issues about it are now happening to me, and could potentially get me kicked out of graduate school for a paper about MY PERSONAL LIFE!!!! MYYYYYY PERSONAL LIFE!!!!!!

Pangram is absurd. We submitted a fully Chat GPT generated text and it says it's fully HUMAN written. The next day, I submitted another one where I created the text and made it as scholarly as possible, and it says it's FULLY AI.

It changed 3 times on the amount of AI content and it is not accurate.

Pangram AI tespit programı, metinler üzerinde tutarlı ve güvenilir sonuçlar vermemektedir. Aynı metne yalnızca bir cümle eklediğimde oranı %0’dan %100’e çıkarması, uzun bir metne eklediğimde ise oranı tamamen farklı değerlendirmesi programın ciddi biçimde hatalı çalıştığını göstermektedir. Ayrıca hiçbir değişiklik yapılmadan tekrar test edildiğinde bile sonuçlar sürekli değişmektedir. Teknik ifadeler, sayısal veriler veya insan tarafından yazılmış nötr cümleler dahi %100 AI olarak işaretlenebilmektedir. Bu nedenle programın yapay zeka tespit algoritması güvenilir değildir ve hatalı pozitif sonuç üretmektedir. Hazırladığım metinler tarafımdan yazılmıştır; söz konusu aracın verdiği sonuçlar bilimsel açıdan geçerli bir ölçüt oluşturmamaktadır.

Really helped me clarify what was ai in papers and what wasn’t thank you

It just does not do what it needs to.

I share 5 sentences it says it's full human written text.
I add 1 one more sentence it's 100% sure that it's AI (all 6).
then I add these 6 sentences to some big text (with hundreds of sentences) it says 40% of text is AI.

I remove it is says 20% of text is AI (text if fully written by person).

Your program finds AI in sentences like
Temperature 100 - 1000 molecules, temperature 200 - 1300 molecues, etc.

It says full AI (100%).

I'd rate it 0/100. Not useful at all, spend my money on it without any result. You say there can be a mistake in 0,01% of texts.
But what you did is full random.

Oh one more important thing. This % of AI text depends on random rerun. You may get 30%, in 1h without changing anything this will be 40%. In next try (next day) this will be 25%. Without any changes.

AI checking AI is not good.

Best regards

Like ashole

Accurate. And provides the insights I need. I love the g docs integration.

I wrote about my problem with signing up but you didn't answer or any help.

it was good and it had a pretty acurate ai detection