Adobe sells fake AI-generated Israel-Hamas war images – then the news ran them as real
(2023/11/09)
- Reference: 1699485730
- News link: https://www.theregister.co.uk/2023/11/08/adobe_ai_israel_hamas_war_pics/
- Source link:
Violent AI-generated photos fictionalizing the ongoing deadly Israel-Hamas conflict are not only being sold via Adobe's stock image library, some news publishers are buying and using the pics in online articles as if they were real.
Last year, the Photoshop titan [1]announced it would host and sell people's images produced using machine-learning tools via its Adobe Stock library. Adobe gives those artists a [2]33 percent cut of the revenue from those pictures, meaning creators could make between 33 cents and $26.40 every time their image is licensed and downloaded.
Generative AI technology has advanced, and become more accessible and easy to use. Anyone can turn to a variety of text-to-image tools to produce realistic-looking images. Some people have now decided to use them to create fake images depicting the war between Israel and Hamas in Gaza, and are selling them on Adobe Stock.
[3]
Although these images are labeled as "generated by AI" in the image library, the disclosure is often not carried forward when it is downloaded and posted elsewhere online, including in news articles run by small-time outlets that reportedly did not mark the snaps as synthetic. Hopefully that's not a sign of things to come.
[4]
Made-up: The AI-generated Adobe stock images used in news articles ... Click to enlarge
An Adobe Stock [5]image titled "Conflict between Israel and Palestine generative AI," that shows black clouds of smoke billowing from buildings, appears to have been published in numerous internet articles as if it were real, Crikey, an Aussie news title, [6]first reported .
A quick search on Adobe Stock shows [7]similar war-related AI-generated images portraying airstrikes, burning vehicles, and destroyed buildings in Gaza.
[8]Microsoft, Meta detail plans to fight election disinformation in 2024
[9]WhatsApp AI happily added guns to chat stickers of Palestinians, but not Israelis
[10]How 'AI watermarking' system pushed by Microsoft and Adobe will and won't work
[11]Adobe to sell AI-generated images on its stock photo platform
Experts have repeatedly warned that AI will be used to spread misinformation online, and it is becoming increasingly difficult to tackle the issue as more and more realistic-looking synthetic content is posted and shared on the internet. Several outfits are trying to, in various ways, watermark AI-generated content so netizens can attempt to tell what's real and fake.
Under the Content Authenticity Initiative, tech and journalism orgs including Adobe, Microsoft, the BBC, and the New York Times are attempting to implement and champion [12]Content Credentials , which uses file metadata to highlight the source of an image, be that human or machine-learning software.
[13]
[14]
Content Credentials, however, hasn't been deployed in practical settings yet, and requires cooperation from social networks, publishers, artists, and application, browser, and generative AI developers to work as hoped.
"Adobe Stock is a marketplace that requires all generative AI content to be labeled as such when submitted for licensing," a spokesperson for the Photoshop giant [15]told PetaPixel.
[16]
"These specific images were labeled as generative AI when they were both submitted and made available for license in line with these requirements. We believe it's important for customers to know what Adobe Stock images were created using generative AI tools.
"Adobe is committed to fighting misinformation, and via the Content Authenticity Initiative, we are working with publishers, camera manufacturers and other stakeholders to advance the adoption of Content Credentials, including in our own products. Content Credentials allows people to see vital context about how a piece of digital content was captured, created or edited including whether AI tools were used in the creation or editing of the digital content."
The Register has asked Adobe for further comment. ®
Get our [17]Tech Resources
[1] https://www.theregister.com/2022/12/06/adobe_ai_images/
[2] https://photutorial.com/adobe-stock-contributor/
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://regmedia.co.uk/2023/11/08/adobe_stock_screenshot.jpg
[5] https://stock.adobe.com/images/conflict-between-israel-and-palestine-generative-ai/606179065?prev_url=detail
[6] https://www.crikey.com.au/2023/11/01/israel-gaza-adobe-artificial-intelligence-images-fake-news/
[7] https://stock.adobe.com/images/gaza-crisis-ceasefire-efforts-amidst-ongoing-attacks/658650709?prev_url=detail
[8] https://www.theregister.com/2023/11/08/microsoft_meta_election_disinformation/
[9] https://www.theregister.com/2023/11/07/whatsapp_ai_stickers_palestinian/
[10] https://www.theregister.com/2023/10/15/microsoft_adobe_ai_watermark/
[11] https://www.theregister.com/2022/12/06/adobe_ai_images/
[12] https://www.theregister.com/2023/10/15/microsoft_adobe_ai_watermark/
[13] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[14] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[15] https://petapixel.com/2023/11/07/adobe-stock-is-selling-ai-generated-images-of-the-israel-hamas-conflict/
[16] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[17] https://whitepapers.theregister.com/
Last year, the Photoshop titan [1]announced it would host and sell people's images produced using machine-learning tools via its Adobe Stock library. Adobe gives those artists a [2]33 percent cut of the revenue from those pictures, meaning creators could make between 33 cents and $26.40 every time their image is licensed and downloaded.
Generative AI technology has advanced, and become more accessible and easy to use. Anyone can turn to a variety of text-to-image tools to produce realistic-looking images. Some people have now decided to use them to create fake images depicting the war between Israel and Hamas in Gaza, and are selling them on Adobe Stock.
[3]
Although these images are labeled as "generated by AI" in the image library, the disclosure is often not carried forward when it is downloaded and posted elsewhere online, including in news articles run by small-time outlets that reportedly did not mark the snaps as synthetic. Hopefully that's not a sign of things to come.
[4]
Made-up: The AI-generated Adobe stock images used in news articles ... Click to enlarge
An Adobe Stock [5]image titled "Conflict between Israel and Palestine generative AI," that shows black clouds of smoke billowing from buildings, appears to have been published in numerous internet articles as if it were real, Crikey, an Aussie news title, [6]first reported .
A quick search on Adobe Stock shows [7]similar war-related AI-generated images portraying airstrikes, burning vehicles, and destroyed buildings in Gaza.
[8]Microsoft, Meta detail plans to fight election disinformation in 2024
[9]WhatsApp AI happily added guns to chat stickers of Palestinians, but not Israelis
[10]How 'AI watermarking' system pushed by Microsoft and Adobe will and won't work
[11]Adobe to sell AI-generated images on its stock photo platform
Experts have repeatedly warned that AI will be used to spread misinformation online, and it is becoming increasingly difficult to tackle the issue as more and more realistic-looking synthetic content is posted and shared on the internet. Several outfits are trying to, in various ways, watermark AI-generated content so netizens can attempt to tell what's real and fake.
Under the Content Authenticity Initiative, tech and journalism orgs including Adobe, Microsoft, the BBC, and the New York Times are attempting to implement and champion [12]Content Credentials , which uses file metadata to highlight the source of an image, be that human or machine-learning software.
[13]
[14]
Content Credentials, however, hasn't been deployed in practical settings yet, and requires cooperation from social networks, publishers, artists, and application, browser, and generative AI developers to work as hoped.
"Adobe Stock is a marketplace that requires all generative AI content to be labeled as such when submitted for licensing," a spokesperson for the Photoshop giant [15]told PetaPixel.
[16]
"These specific images were labeled as generative AI when they were both submitted and made available for license in line with these requirements. We believe it's important for customers to know what Adobe Stock images were created using generative AI tools.
"Adobe is committed to fighting misinformation, and via the Content Authenticity Initiative, we are working with publishers, camera manufacturers and other stakeholders to advance the adoption of Content Credentials, including in our own products. Content Credentials allows people to see vital context about how a piece of digital content was captured, created or edited including whether AI tools were used in the creation or editing of the digital content."
The Register has asked Adobe for further comment. ®
Get our [17]Tech Resources
[1] https://www.theregister.com/2022/12/06/adobe_ai_images/
[2] https://photutorial.com/adobe-stock-contributor/
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://regmedia.co.uk/2023/11/08/adobe_stock_screenshot.jpg
[5] https://stock.adobe.com/images/conflict-between-israel-and-palestine-generative-ai/606179065?prev_url=detail
[6] https://www.crikey.com.au/2023/11/01/israel-gaza-adobe-artificial-intelligence-images-fake-news/
[7] https://stock.adobe.com/images/gaza-crisis-ceasefire-efforts-amidst-ongoing-attacks/658650709?prev_url=detail
[8] https://www.theregister.com/2023/11/08/microsoft_meta_election_disinformation/
[9] https://www.theregister.com/2023/11/07/whatsapp_ai_stickers_palestinian/
[10] https://www.theregister.com/2023/10/15/microsoft_adobe_ai_watermark/
[11] https://www.theregister.com/2022/12/06/adobe_ai_images/
[12] https://www.theregister.com/2023/10/15/microsoft_adobe_ai_watermark/
[13] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[14] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[15] https://petapixel.com/2023/11/07/adobe-stock-is-selling-ai-generated-images-of-the-israel-hamas-conflict/
[16] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZUxndjnd1lrq0iptUOGsCQAAAAQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[17] https://whitepapers.theregister.com/
Re: Metadata... yea, that'll solve the problem
elsergiovolador
And this also requires the readers to read the fine print, which I find many people don't do. It seems to be a lost art these days.
The problem is that real journalism today virtually doesn't exist and reading articles to the last word is a waste of time. "Journalists" have an agenda that is set by the shareholders / owners / spooks and government and they have to keep things in line if they want to keep their jobs or life.
In the grand scheme of things, these "news" are irrelevant to most people anyway. You can't change what's going on, so you might as well free yourself from reading about it.
Dimmer
We recently had a vote on new constitution amendments. If you read it, it will not raise taxes and it will make the world a better place.
One of the amendments was about water for a specific county. After it passed, they then told everyone that they voted to create a new tax entity called a water board, and it was for 6 counties and they will be soon footing the bill so real estate developers would not have to.
I studied this before voting and totally missed it because there was no fine print.
There should be a law that requires them to put in plain words what the amendment is for or it is void.
Metadata... yea, that'll solve the problem
Curious how this "metadata" survives a social media (re)post or reproduction in physical media or on other online publications' websites.
Of course, conscientious publishers show their sources for images - but that doesn't translate in social media well. One just has to consider all the "sponsored" posts by "influencers" that neglected that tidbit.
By stating the source is "AI Generated"... kinda ruins the point of using the photo for a real article about a real event. No one sees "artist's rendition" from either of the Iraq wars. Court room illustrations stand as a counter-example... but those are clearly not life-like.
And this also requires the readers to read the fine print, which I find many people don't do. It seems to be a lost art these days.
Hell, even this website's articles seems to have recently added a bold banner for older articles (or at least I just started to notice it). Perhaps because people find it difficult to care to look at an article's date of publishing, which tends not to be in the largest of fonts.