AI map fails show how unintelligent AI can be
AI's hopeless map generation shows it is matching patterns, rather than understanding all the information from the often stolen data which it has copied.
Making a map requires an understanding of the underlying data in the map, which the 'Artificial Intelligence'/Pattern Recognition doesn't really have, while it rearranges and reproduces bits of unlicensed copyrighted material. So it continues to make all sorts of shocking map errors which no human would ever make.
Sometimes it makes up country names, or repeats them in wrong places, or adds nonsensical text, or puts whole countries in the wrong area, or does all sorts of 'hallucinations' that make no sense at all.
What's weird is some parts are correct (doubtless lifted straight from another copyrighted map), which can trick you into thinking the image is right, while other parts can be a total mess of inaccuracy. Like the AI map at the top of this page: Canada looks almost correct, I'm guessing because it was copied from some other map. While you can kind of see how it got 'UNITES' from 'UNITED STATES,' though that's still a huge error. But Europe is an absurd disaster with nonsensical borders and countries labelled 'ISIAE, MAIJJA and IRELION' ...what the?!
*This is a cartographer's opinion on AI maps:
“These AI systems aren't making maps. They're making pictures of maps. They've seen millions of map images and they're very good at generating something that looks like one, but they have zero understanding of what they're producing.”
Quoted from 'AI Makes Maps Now. That's a Problem'
https://www.youtube.com/watch?v=uB8XUxkiN0o
*'Bad AI Maps' from a site called 'Map of the Week':
https://mapoftheweek.blogspot.com/2026/03/bad-ai-maps-part-1.html
https://mapoftheweek.blogspot.com/2026/03/bad-ai-maps-part-2.html
*Video compilations of wrong AI maps with many mistakes:
AI Maps Are Still So Bad
https://www.youtube.com/watch?v=W89YWBhFaww
AI Maps That Make me Want to End It
https://www.youtube.com/watch?v=b7USVYPBbNw
*AI couldn't even make a simple map labelling 6 countries on a map of Africa, which caused trouble for the U.S. State Department:
'US government map of Africa mislabels every country at global conference'
https://www.channelnewsasia.com/world/us-africa-government-map-mislabel-country-ai-6288891
*Here is an objective test with someone prompting different AI programs with the same map generation queries, and producing many problems:
I Asked AI to Draw a Map...
https://www.youtube.com/watch?v=YLoJITf6F24
*The source for the image at the top of this page, showing the full world map, including locations of 'countries' such as INICEP, ZOPHON, GEELONA and CLUIEM SROOS:
'Can we take a minute to make fun of how inaccurate this AI generated map is'
https://www.reddit.com/r/antiai/comments/1r1inin/can_we_take_a_minute_to_make_fun_of_how/
*More error filled AI maps on Reddit:
'AI slop maps'
https://www.reddit.com/r/mapporncirclejerk/comments/1v7ldxo/ai_slop_maps/
'The World Map, according to AI'
https://www.reddit.com/r/geographymemes/comments/1jzpcs9/the_world_map_according_to_ai/
'AI map of the world using famous landmarks'
https://www.reddit.com/r/aimapgore/comments/1tpo4gk/ai_map_of_the_world_using_famous_landmarks/
'Found this AI slop map... [of the British Empire]'
At first glance the AI maps sometimes almost look right, because they copy the style of an original map, but the more you look, the more errors you spot.
This problem, so well illustrated with maps, is the same with all Generative AI content: it combines bits of stolen copyrighted material of high quality, with generative material of variable quality, which can be terrible or wrong. So Generative AI is this weird mix that just feels awful, and you never know when the quality is going to dip: it's about as variable as all the copyrighted work which it stole from to make the new material.
In contrast, human produced work tends to be about the same quality: from professional level publishers at the top, then amateur level in the middle and maybe rushed work or kids drawings and so on at the bottom. But they were all vacuumed up by the AI which has no respect for the copyrights of anyone: professionals, amateurs, kids - they stole them all!
Since AI companies stole everything, the quality of their work goes up and down like a rollercoaster, in the same work. Which of course, is why they shouldn't be allowed to steal copyrighted material, and why they should release all data on copyrighted work they've copied.
***
*We all make mistakes. (The AI more than others) But on the last update I said 'doom scrolling' when I meant 'doom trolling.' Whoops.
This is the Doom Trolling I meant:
'"Doom trolling" is Cal Newport's term for AI companies warning us their own products might destroy civilization.'
*He explains this doom trolling further here:
Dear AI Companies: Stop the “Doom Trolling”
https://www.youtube.com/watch?v=aHxwZEXHwps
*The latest round of AI companies doom trolling was debunked by Ed Zitron in
'Is AI The Terminator Or The Big Short?'
(The Big Short is a film about the 'housing bubble'/2008 financial crisis - which I haven't seen, I had to look that up)
https://www.youtube.com/watch?v=1qlGoRI94rQ
*Ed's a reporter investigating the AI industry and has brought up numerous problems in many interviews available on Youtube, which often get hundreds of thousands of views.
One of his recent concerns is that a large number of AI data centers are unfinished projects, while the computer chips are being hoarded by AI companies in warehouses, the same computer parts taken away from everyday people. I think that's what he's saying, some of this business/computer information is too technical for me, but you can find out more here:
'The Tech Report,' AI Bubble: ‘This could humiliate the largest companies in the world’
' "Half of Nvidia’s revenue could literally be sitting in warehouses."... why the AI industry is continuing to buy AI chips despite reports that significant amounts of GPUs [Graphics Processing Units] are going uninstalled.'
https://www.youtube.com/watch?v=8C-J2sRBMkQ
That's quite a technical interview which goes beyond my understanding. But luckily Ed's been on mainstream news lately now too, such as in this shorter 6 minute interview here:
Ed Zitron on ABC News talking about actually regulating AI and the truth behind its economics
https://www.youtube.com/watch?v=lv7AWjqVRhQ
On Ed's website https://www.wheresyoured.at/ he points out a range of financial problems which the AI industry are facing. In the middle of one detailed article that was going over my head, was this simple explanation of the costs:
"...big tech has a few issues with the AI buildout:
* AI chips are extremely expensive.
* AI data centers take a great deal of time to build and energize.
* AI services are expensive to run.
* AI services do not appear to generate much revenue."
*These costs are also shown graphically as a $900 billion loss at the site 'Is AI profitable yet?'
So they lose all this money while they steal copyrighted work, ruin society with slop, and pollute the environment with data centres.
***
*Update on case numbers involving 'Artificial Intelligence' in America:
Copyright infringement cases against AI companies: 145 (135 active)
Civil law suits against AI companies: 124
Data centre lawsuits: 52
Reference: https://chatgptiseatingtheworld.com/
*Along with the lawsuits, other actions against data centre construction continue.
*Several U.S. states are in the process of making moritoriums against their construction, which are described here:
National Conference of State Legislatures, 'Which States Are Banning Data Centers?'
https://www.ncsl.org/fiscal/which-states-are-banning-data-centers
*The 'Data Center Moratoriums' page records the following numbers for U.S. states:
Active Bans / Moratoriums: 20
Legislation Advancing: 7
Under Discussion: 15
Source: https://www.datacenterbans.com/
*Local level data centre bans are described in more detail on other pages, including the 'Data Center Moratoriums Tracker,' which says there are 41 states with local bans, totalling at least 294 moratoriums adopted in the first half of 2026.
Source: https://programs.com/resources/data-center-moratoriums-tracker/
*An interactive map of AI data centres including a place to report them is here:
https://www.brockovichdatacenter.com/
*Data centres are not just responsible for environmental pollution of air, land and water, but also cause heavy power and water usage. New AI data centres may use more power than the grid can handle, taking up land and causing high electric bills, described here:
Data Centers Are Breaking The Power Grid
https://www.youtube.com/watch?v=X71le3av1So
*Another site trying to stop data centres I discovered recently:
*At change.org are a large number of petitions against data centres, with a group of U.S. data centre protests in the movement 'Fighting Data Center Development.'
This movement has at the time of writing, 1.3 million supporters and 175 separate petitions!
https://www.change.org/m/halting-data-center-development
One petition received 600k signatures:
Nashville Zoo Says No to Proposed Data Center
https://www.change.org/p/nashville-zoo-says-no-to-proposed-data-center
*The British data centre petition movement, 'Opposition to Data Centres in the UK' has 150k supporters and 30 separate petitions:
https://www.change.org/m/opposition-to-data-centres-in-the-uk
*Another movement covers data centres in Australia/New Zealand:
'Halt data centre development Australia/NZ', with 44 separate petitions:
https://www.change.org/m/no-data-centres-in-australia-nz
*Also several large petitions are against data centers in the Philippines, with more than half a million signatures combined from these two petitions:
Stop the Signing of the Pax Silica Treaty in November Until We're Assured of Safeguards
Reject #NakakaPAX... SILICA in the Philippines! [written in Filipino]
https://www.change.org/p/reject-nakakapaxshit-silica-in-the-philippines
*I just hope they don't use AI to make more maps of the Philippines, or PHILIPRINES, as it's misspelled in 'Map of the Philippines according to ChatGPT,' while the country also seems to have annexed bits of Indonesia and Malaysia:
*At least PHILIPRINES is better than BMOHRMES as the Philippines is known at the 'AI World Map,' while Malaysia is called KEAIJONA, Spain is known as ROBIRIEAL and Germany appears as SGIRAMM, amongst numerous errors here:
https://www.reddit.com/r/mapporncirclejerk/comments/1m05jgr/ai_world_map/
And that map has a country called DANG, which is exactly what you think when you see AI maps.
Dang!
*On a lighter note is this parody of advertisements for AI companies:
AI Companies Trying to Make Normal Ads
https://www.youtube.com/watch?v=zwE3eF3Tg1M
***
AN ASTONISHING THEFT OF UNPRECEDENTED PROPORTIONS
A COMPLETE MOCKERY OF THE IDEA OF FAIR USE
In the petition I talked about how AI companies copying of copyrighted content is not Fair Use. Documents released a few weeks ago from New York Times v. OpenAI/Microsoft have shown that people within the tech companies felt the same way.
Microsoft’s director of applied science wrote:
"Millions of people around the world will soon consider large models ‘hoovering up’ all their work to be an astonishing theft of unprecedented proportions.” *
Yep, he's right about that. He goes on to say this would be the "largest theft of labor in human history,” which would "make a complete mockery of the idea of ‘fair use.’" *
*(Note: this is an individual perspective which does not reflect the company's views, according to Microsoft.)
Reference: News Plaintiffs' Combined Summary Judgement Brief
https://storage.courtlistener.com/recap/gov.uscourts.nysd.640396/gov.uscourts.nysd.640396.1977.1.pdf
Other comments from released documents are described in an article by the Nieman Foundation at Harvard:
“An astonishing theft of unprecedented proportions”: Court records show what Microsoft and OpenAI actually thought about AI training.
*The U.S. Copyright Office also took the view that the use of copyrighted material for AI training is often not Fair Use, in the conclusion of its May 2025 report:
"making commercial use of vast troves of copyrighted works to produce expressive content that competes with them in existing markets, especially where this is accomplished through illegal access, goes beyond established fair use boundaries."
Reference: 'Copyright and Artificial Intelligence, Part 3: Generative AI training'
But that tech guy said it best,
AN ASTONISHING THEFT OF UNPRECEDENTED PROPORTIONS
A COMPLETE MOCKERY OF THE IDEA OF FAIR USE