Новини

NEW YORK — Artists under siege by artificial intelligence that studies their work and then replicates their styles, have teamed with university researchers to stymie such copycat activity.

U.S. illustrator Paloma McClain went into defense mode after learning that several AI models had been trained using her art, with no credit or compensation sent her way.

“It bothered me,” McClain told AFP.

“I believe truly meaningful technological advancement is done ethically and elevates all people instead of functioning at the expense of others,” she said.

The artist turned to free software called Glaze created by researchers at the University of Chicago.

Glaze essentially outthinks AI models when it comes to how they train, tweaking pixels in ways that are indiscernible to human viewers but which make a digitized piece of art appear dramatically different to AI.

“We’re basically providing technical tools to help protect human creators against invasive and abusive AI models,” said Ben Zhao, a professor of computer science on the Glaze team.

Created in just four months, Glaze spun off technology used to disrupt facial recognition systems.

“We were working at super-fast speed because we knew the problem was serious,” Zhao said of rushing to defend artists from software imitators. “A lot of people were in pain.”

Generative AI giants have agreements to use data for training in some cases, but the majority of digital images, audio, and text used to shape the way supersmart software thinks has been scraped from the internet without explicit consent.

Since its release in March, Glaze has been downloaded more than 1.6 million times, according to Zhao.

Zhao’s team is working on a Glaze enhancement called Nightshade that notches up defenses by confusing AI, say by getting it to interpret a dog as a cat.

“I believe Nightshade will have a noticeable effect if enough artists use it and put enough poisoned images into the wild,” McClain said, meaning they would be easily available online.

“According to Nightshade’s research, it wouldn’t take as many poisoned images as one might think,” she said.

Zhao’s team has been approached by several companies that want to use Nightshade, according to the Chicago academic.

“The goal is for people to be able to protect their content, whether it’s individual artists or companies with a lot of intellectual property,” Zhao said.

Viva Voce

A startup called Spawning has developed Kudurru software that detects attempts to harvest large numbers of images from an online venue.

An artist can then block access or send images that don’t match what is being requested, tainting the pool of data being used to teach AI what is what, according to Spawning co-founder Jordan Meyer.

More than 1,000 websites have been integrated into the Kudurru network.

Spawning has also launched haveibeentrained.com, a website that features an online tool for finding out whether digitized works have been fed into an AI model and allow artists to opt out of such use in the future.

As defenses ramp up for images, researchers at Washington University in Missouri have developed AntiFake software to thwart AI copying voices.

AntiFake enriches digital recordings of people speaking, adding noises inaudible to people but which make it “impossible to synthesize a human voice,” said Zhiyuan Yu, the Ph.D. student behind the project.

The program aims to go beyond just stopping unauthorized training of AI to preventing the creation of “deepfakes” — bogus soundtracks or videos of celebrities, politicians, relatives, or others showing them doing or saying something they didn’t.

A popular podcast recently reached out to the AntiFake team for help stopping its productions from being hijacked, according to Zhiyuan Yu.

The freely available software has so far been used for recordings of people speaking, but could also be applied to songs, the researcher said.

“The best solution would be a world in which all data used for AI is subject to consent and payment,” Meyer contended. “We hope to push developers in this direction.”
…

HERZLIYA, Israel — Nearly 7,000 miles away in Portland, Oregon, venture capitalist George Djuric said he was compelled to visit Israel during the country’s war with Palestinian militant group Hamas and to pledge support for the high-tech sector.

Djuric, chief technology officer at yVentures who arrived in the United States as a 3-year-old refugee from Bosnia during the Bosnian war in the mid-1990s, this week joined some 70 other U.S. tech executives and investors on a trip to Israel.

“Coming here is a chance to stand in solidarity with Israel and also support the tech ecosystem, which is the world’s second largest after Silicon Valley,” he said. “As a technology fund, it makes sense for us to be here.”

Although not Jewish, Djuric said he was drawn to Israel by the state’s resiliency and as someone whose family’s views were shaped by war.

“I was horrified by what happened on October 7 and I was equally horrified the next day when I saw people demonstrating in support of what happened,” he said, referring to the October 7 attack on Israel launched by Hamas.

Investors and analysts had predicted the conflict with the Palestinians would derail a fragile recovery in high-tech, which accounts for more than half of Israel’s exports and nearly a fifth of its overall economic output.

Funding had already dropped sharply amid a global slowdown and a divisive government judicial overhaul when the war took its toll on the economy. Growth, on pace for a 3.4% clip this year, has fallen to an expected 2% with the outlook at least as grim.

At least 15% of the tech workforce has been called up for military reserve duty.

Yet, even as the war rages, tech funding deals are still getting done, albeit at a slower pace. Startups have raised more than $6 billion in 2023 compared with $16 billion in 2022.

On Tuesday, ScaleOps, a startup specializing in cloud resource management, announced a $21.5 million funding round. Last week, cyber startup Zero Networks, which prevents attackers from spreading in corporate networks, raised $20 million.

‘Long-term bullish on Israel’

Ron Miasnik, of Bain Capital Ventures who co-organized the delegation, said he had expected Israeli startups to go on drawing large sums. He said he believed the country’s economy would ultimately bounce back.

“It doesn’t matter to us whether the economic rebound takes three months, six months, nine months or 12 months,” he said. “We’re long-term bullish on Israel.”

Miasnik said the idea of the trip emerged from watching other solidarity groups, such as religious ones. “We felt the (U.S.) tech and the venture capital community, which is so heavily integrated within Israel, was missing,” he said.

Initially, it was supposed to be just 15 people but, he said, hundreds of people showed interest. They included CEOs and senior executives of U.S.-based tech and VC funds from Meetup.com, Apollo, TPG, Susquehanna Growth Equity, Mastercard, John Deere and Harvard University’s endowment investment fund.

In addition to meeting local investors and startups, they met Israeli leaders and families of hostages still held captive in Gaza and toured border towns hit by the October 7 attack.

Bain has a number of investments in Israel, including Redis Labs, in which the fund has invested more than $100 million, and cybersecurity firm Armis, and Miasnik said he was seeking to add more Israeli cybersecurity startups to its portfolio.

Similarly, Danny Schultz, managing director of New York-based Gotham Ventures said he was looking to invest in 10 to 20 Israeli growth stage startups, mainly in fintech, in the next three to five years.

“At the point that Israeli CEOs need more capital, they also need relationships across the ocean in the U.S. and Europe to really help build their companies,” he said.

Joy Marcus co-founded a new VC fund called The 98 and only invests in “women-led technology businesses that are disrupting industry.”

“I am tortured by the war. … So I am here to support Israel first and foremost,” she said. “And I am also very interested in investing in some Israeli women.”
…

…

MOSCOW — The head of a company that makes navigation systems for Russia’s space program was arrested in Moscow and charged with major fraud, state media reported Friday.

TASS news agency quoted an unidentified law enforcement official as saying that Yevgeny Fomichev had been interrogated and charged with large-scale fraud, which carries a prison term of up to 10 years and a fine of 1 million rubles ($10,972).

TASS said Moscow’s Basmanny District Court, which often handles high-profile cases, ordered Fomichev to be held in pretrial detention until Feb. 21 at the request of Russia’s Investigative Committee, which deals with serious crimes.

Fomichev is head of NPP Geophysics-Cosmos, a company whose website says it manufactures “optical electronic orientation and navigation devices for spacecraft.” It says that almost all Russian spacecraft use its equipment.

The website includes a nine-page anti-corruption policy that says management has a key role in creating a culture of zero-tolerance toward corruption.

Russia’s space program suffered a huge setback in August when its Luna-25 spacecraft smashed into the surface of the moon while attempting to land there.

An investigation blamed a malfunction in an onboard control unit for the failure of Russia’s first moon mission in 47 years.
…