When President Biden signed a landmark executive order on AI in October, he directed the government to develop standards for companies to follow in watermarking their images. ...
“Watermarking will definitely help,” Dekens said. But “it’s certainly not a waterproof solution, because anything that’s digitally pieced together can be hacked or spoofed or altered,” he said. ...
Labeling real images
On top of watermarking AI images, the tech industry has begun talking about labeling real images as well, layering data into each pixel right when a photo is taken by a camera to provide a record of what the industry calls its “provenance.” ...
“It’s dangerous to believe there are actual solutions against malignant attackers,” said Vivien Chappelier, head of research and development at Imatag, a start-up that helps companies and news organizations put watermarks and labels on real images to ensure they aren’t misused. But making it harder to accidentally spread fake images or giving people more context into what they’re seeing online is still helpful. ...
Detection software
Some companies, including Reality Defender and Deep Media, have built tools that detect deepfakes based on the foundational technology used by AI image generators.
By showing tens of millions of images labeled as fake or real to an AI algorithm, the model begins to be able to distinguish between the two, building an internal “understanding” of what elements might give away an image as fake. Images are run through this model, and if it detects those elements, it will pronounce that the image is AI-generated. ...
There are other things to look for, too, such as whether a person has a vein visible in the anatomically correct place, said Ben Colman, founder of Reality Defender. “You’re either a deepfake or a vampire,” he said. ...
“If the problem is hard today, it will be much harder next year,” said Feizi, the University of Maryland researcher. “It will be almost impossible in five years.”
Assume it’s fake
... “Assume nothing, believe no one and nothing, and doubt everything,” said Dekens, the open-source investigations researcher. “If you’re in doubt, just assume it’s fake.” ...
See the full story here: https://www.washingtonpost.com/technology/2024/04/05/ai-deepfakes-detection
... Today’s large language models (LLMs) have learnt to recognise patterns but don’t understand the underlying concepts. They will therefore always produce silly errors, says Marcus. ...
We don’t do very much to adapt to AI (or climate change) because we can’t quite believe upheaval until it arrives....
Even tech optimists find themselves caught out. Meta’s head of AI, Yann LeCun, told world leaders on February 13 that a text-to-video generating AI service was not possible: “Basically we don’t know how to do this.” A few days later, OpenAI revealed its text-to-video model, Sora. If you loudly say that AI will never be able to do something, there’s the chance that someone in Silicon Valley is laughing. ...
Like crypto, AI has identifiable flaws. LLMs such as OpenAI’s can’t digest all human knowledge. They are trained on sets of available data — words, images and audio, but not the direct interaction with the physical world. Even if you pump in more data, can you address the limitations? ...
“The burden of proof lies with the people making the extraordinary claims . . . No one is saying AI is hype, we’re saying that your claims of AI are hype.” ...
Gary Marcus suggests performance may get worse: LLMs produce untrustworthy output, which is then sucked back into other LLMs. The models become permanently contaminated. Scientific journals’ peer-review processes will be overwhelmed, “leading to a precipitous drop in reputation”, Marcus wrote recently. ...
(Rasenberger adds that ChatGPT and others would face restrictions on how they used copyrighted data: they would not, for example, be able to provide text in the style of a certain author.) ...
Recently scepticism has got a bad name, because of how easily its followers have veered into conspiracism: doubting credible information about climate, vaccines and Ukraine. AI scepticism has so far avoided this fate. ...
See the full story here: https://www.ft.com/content/648228e7-11eb-4e1a-b0d5-e65a638e6135
... Right now, everyone’s searching for “the iPhone of AI,” but we’re not getting that anytime soon. We might not get it ever, for that matter, because the promise of AI is that it doesn’t require a certain kind of perfected interface — it doesn’t require any interface at all. What we’re going to get instead are the Razr, the Chocolate, the Treo, the Pearl, the N-Gage, and the Sidekick of AI. It’s going to be chaos, and it’s going to be great.
See the full story here; https://www.theverge.com/24117865/ai-gadget-era-humane-rabbit-brilliant-meta
“This is not real audio/video of Drake or Queen Elizabeth II. All video and speech was computer generated to help others learn about math, physics, and engineering,” @onlocklearning writes in the caption for a video of the rapper and the late monarch explaining trigonometry.
And according to the comments on the platform, it appears that these celebrity deepfakes are genuinely helping young viewers understand mathematical theories.
“I learned something in one minute [that] my teacher would have taken a whole class or two to teach us,” a TikTok user comments. ...
“They may pay attention to these for a very short period of time, but the question is how much do they really learn from these? Does this really promote deep learning?”
See the full story here: https://petapixel.com/2024/04/02/a-deepfake-taylor-swift-is-teaching-math-to-kids-on-tiktok/?utm_source=substack&utm_medium=email
Researchers from The Barcelona Institute of Science and Technology’s ICFO–Institut de Ciències Fotòniques have developed the first ever semi-transparent image sensor.
This sensor comprises an 8×8 array of semi-transparent photodetectors and electrodes disposed on a fully transparent substrate. Each pixel in the array is of size 60 x 140 μm and has an optical transparency of 85-95 percent, according to the study.
These photodetectors capture light while allowing a considerable portion of it to pass through— a necessary trait for applications where transparency is essential, such as in smart displays on AR and VR devices.
Additionally, this sensor sports a design that balances light capture with visibility, making it suitable for applications requiring both sensing capabilities and transparency. ...
See the full story here: https://interestingengineering.com/innovation/first-semi-transparent-image-sensor
Big tech has substantial influence over the direction of R&D in the US. According to the National Science Foundation and the Congressional Research Service, US business R&D spending dwarfs domestic Federal or state government spending on research and development. ...
Of course, business R&D spending focuses mainly on development–76 percent versus 14 percent on applied research and seven percent on basic research. ...
Private sector R&D is quite concentrated in a handful of companies with dominant market shares and capitalizations. From a tech perspective, these companies might claim to be on the bleeding edge of innovation, but are all protecting cash cow legacy stacks and a legacy architecture mindset. The innovations they’re promoting in any given year are the bright and shiny objects, the VR goggles, the smart watches, the cars and the rockets. Meanwhile, what could be substantive improvement in infrastructure receives a fraction of their investment. Public sector R&D hasn’t been filling in these gaps. ...
I hope the US, which has seemed to be relatively leaderless on the R&D front over the past decade, can emulate more visionary European efforts like this one from the Swiss in the near future.
See the full story here: https://www.datasciencecentral.com/rd-misdirection-and-the-circuitous-us-path-to-artificial-general-intelligence/
Talent agency WME has inked a partnership with Loti, a Seattle-based firm that specializes in software used to flag unauthorized content posted on the internet that includes clients’ likenesses. The company, which has 25 employees, then quickly sends requests to online platforms to have those infringing photos and videos removed.
Financial details of the deal were not disclosed. ...
Loti co-founder Luke Arrigoni launched the startup about a year and a half ago. He previously ran an artificial intelligence firm called Arricor AI and before that was a data scientist at Creative Artists Agency, WME’s main rival.
Arrigoni said Loti began working with WME about four or five months ago. WME clients give Loti a few photos of themselves from different angles. They also record short audio clips that are then used to help identify unauthorized content. Loti’s software searches the web and reports back to the clients about these unauthorized images and sends takedown requests to the platforms.
“There’s this kind of growing feeling that this is an impossible problem,” Arrigoni said. “There’s this almost adage now where people say, ‘Once it’s on the internet, it’s on the internet forever.’ Our whole company dispels that myth.” ...
See the full story here: https://www.latimes.com/entertainment-arts/business/story/2024-04-02/wme-loti-deepfakes-artificial-intelligence
Apple researchers have developed a new artificial intelligence system that can understand ambiguous references to on-screen entities as well as conversational and background context, enabling more natural interactions with voice assistants, according to a paper published on Friday.
The system, called ReALM (Reference Resolution As Language Modeling), leverages large language models to convert the complex task of reference resolution — including understanding references to visual elements on a screen — into a pure language modeling problem. This allows ReALM to achieve substantial performance gains compared to existing methods. ...
See the full story here: https://venturebeat.com/ai/apple-researchers-develop-ai-that-can-see-and-understand-screen-context/
On Tuesday, the Artist Rights Alliance (ARA) announced an open letter critical of AI signed by over 200 musical artists, including Pearl Jam, Nicki Minaj, Billie Eilish, Stevie Wonder, Elvis Costello, and the estate of Frank Sinatra. In the letter, the artists call on AI developers, technology companies, platforms, and digital music services to stop using AI to "infringe upon and devalue the rights of human artists." A tweet from the ARA added that AI poses an "existential threat" to their art. ...
In considering AI's potential impact on music, it's instructive to remember historical instances where tech innovations initially sparked concern among artists. For instance, the introduction of synthesizers in the 1960s and 1970s and the advent of digital sampling in the 1980s both faced scrutiny and fear from parts of the music community, but the music industry eventually adjusted.
While we've seen fear of the unknown related to AI going around quite a bit for the past year, it's possible that AI tools will be integrated into the music production process like any other music production tool or technique that came before. It's also possible that even if that kind of integration comes to pass, some artists will still get hurt along the way—and the ARA wants to speak out about it before the technology progresses further. ...
See the full story here: https://arstechnica.com/information-technology/2024/04/billie-eilish-pearl-jam-200-artists-say-ai-poses-existential-threat-to-their-livelihoods/?mc_cid=913213710c&mc_eid=116e9f337b
