European Union negotiators clinched a deal Friday on the world’s first comprehensive artificial intelligence rules, paving the way for legal oversight of technology used in popular generative AI services such as ChatGPT that have promised to transform everyday life and spurred warnings of existential dangers to humanity. 

Negotiators from the European Parliament and the bloc’s 27 member countries overcame big differences on controversial points, including generative AI and police use of facial recognition surveillance, to sign a tentative political agreement for the Artificial Intelligence Act. 

“Deal!” tweeted European Commissioner Thierry Breton, just before midnight. “The EU becomes the very first continent to set clear rules for the use of AI.” 

The result came after marathon closed-door talks this week, with the initial session lasting 22 hours before a second round kicked off Friday morning. 

Officials were under the gun to secure a political victory for the flagship legislation but were expected to leave the door open to further talks to work out the fine print, likely to bring more backroom lobbying. 

Out front

The EU took an early lead in the global race to draw up AI guardrails when it unveiled the first draft of its rulebook in 2021. The recent boom in generative AI, however, sent European officials scrambling to update a proposal poised to serve as a blueprint for the world. 

The European Parliament will still need to vote on it early next year, but with the deal done, that’s a formality, Brando Benifei, an Italian lawmaker co-leading the body’s negotiating efforts, told The Associated Press late Friday. 

“It’s very, very good,” he said by text message after being asked if it included everything he wanted. “Obviously we had to accept some compromises but overall very good.”  

The eventual law wouldn’t fully take effect until 2025 at the earliest and threatens stiff financial penalties for violations of up to $38 million (35 million euros) or 7% of a company’s global turnover. 

Generative AI systems like OpenAI’s ChatGPT have exploded into the world’s consciousness, dazzling users with the ability to produce humanlike text, photos and songs but raising fears about the risks the rapidly developing technology poses to jobs, privacy and copyright protection, and even human life itself. 

Now, the U.S., U.K., China and global coalitions like the Group of Seven major democracies have jumped in with their own proposals to regulate AI, though they’re still catching up to Europe. 

‘A powerful example’

Strong and comprehensive regulation from the EU “can set a powerful example for many governments considering regulation,” said Anu Bradford, a Columbia Law School professor who’s an expert on EU and digital regulation. Other countries “may not copy every provision but will likely emulate many aspects of it.” 

AI companies that will have to obey the EU’s rules will also likely extend some of those obligations to markets outside the continent, she said. “After all, it is not efficient to retrain separate models for different markets,” she said. 

Others are worried that the agreement was rushed through. 

“Today’s political deal marks the beginning of important and necessary technical work on crucial details of the AI Act, which are still missing,” said Daniel Friedlaender, head of the European office of the Computer and Communications Industry Association, a tech industry lobby group. 

The AI Act was originally designed to mitigate the dangers from specific AI functions based on their level of risk, from low to unacceptable. But lawmakers pushed to expand it to foundation models, the advanced systems that underpin general purpose AI services like ChatGPT and Google’s Bard chatbot. 

Foundation models looked set to be one of the biggest sticking points for Europe. However, negotiators reached a tentative compromise early in the talks, despite opposition led by France, which called instead for self-regulation to help homegrown European generative AI companies competing with big U.S. rivals, including OpenAI’s backer Microsoft. 

Also known as large language models, these systems are trained on vast troves of written works and images scraped off the internet. They give generative AI systems the ability to create something new, unlike traditional AI, which processes data and completes tasks using predetermined rules. 

Under the deal, the most advanced foundation models that pose the biggest “systemic risks” will get extra scrutiny, including requirements to disclose more information, such as how much computing power was used to train the systems. 

Elevation of threats

Researchers have warned that these powerful foundation models, built by a handful of big tech companies, could be used to supercharge online disinformation and manipulation, cyberattacks or creation of bioweapons. 

Rights groups also caution that the lack of transparency about data used to train the models poses risks to daily life because they act as basic structures for software developers building AI-powered services. 

What became the thorniest topic was AI-powered facial recognition surveillance systems, and negotiators found a compromise after intensive bargaining. 

European lawmakers wanted a full ban on public use of facial scanning and other “remote biometric identification” systems because of privacy concerns, while governments of member countries wanted exemptions so law enforcement could use them to tackle serious crimes like child sexual exploitation or terrorist attacks. 

Civil society groups were more skeptical. 

“Whatever the victories may have been in these final negotiations, the fact remains that huge flaws will remain in this final text,” said Daniel Leufer, a senior policy analyst at the digital rights group Access Now. Along with the law enforcement exemptions, he also cited a lack of protection for AI systems used in migration and border control, and “big gaps in the bans on the most dangerous AI systems.” 
…

A new optical ground station has been built by the Australian National University to help the U.S. space agency, NASA, and others explore space and safely reach Mars.

The Australian team has developed a new type of space communication using lasers.

Researchers say the system will allow them to connect with satellites and NASA-crewed missions beyond low-Earth orbit.

The project is supported by the Australian Space Agency’s Moon to Mars initiative.

The Australian National University Quantum Optical Ground Station is based at the Mount Stromlo Observatory, near Canberra.

It is a powerful telescope that will support high-speed advanced communications with satellites orbiting at distances from low-Earth orbit to the moon.

Kate Ferguson, associate director for strategic projects at the Australian National University Institute for Space, told VOA current communication systems relying on radio frequencies can be slow and cumbersome.

“I am sure some of us remember the grainy pictures that we got of the moon landing that came from the Apollo era,” Ferguson said. “So, again the current radio frequency systems, they have these much slower data rates and especially over really long distances.  For space exploration those become very slow but with optical communications we will be able to increase the rate of that communication.”

She said the new system, based on powerful lasers that are invisible to the naked eye, will transform communications in space.

“What we are aiming to do is to be able to receive high-definition video from future crewed missions. Not only will that be great for us here on Earth, seeing what is happening with the astronauts on these types of missions, but it will improve the connectivity between those missions,” she said. “And what we are doing here is optical communication, which uses laser beams to communicate and these offer much higher speeds and increased security over the current systems and this is really important for us to be getting that data down and being able to use it here on Earth.”

Scientists say the Australian-developed systems will be compatible with NASA missions.

They say the laser-based technology will improve astronauts’ ability to connect with Earth from the moon and also allow high-definition video to be sent from the moon and Mars.

NASA has said previously that astronauts could be sent on a mission to the red planet by the mid-to-late 2030s.

 
…

Facebook and Instagram fail to protect underage users from exposure to child sexual abuse material and let adults solicit pornographic imagery from them, New Mexico’s attorney general alleges in a lawsuit that follows an undercover online investigation.

“Our investigation into Meta’s social media platforms demonstrates that they are not safe spaces for children but rather prime locations for predators to trade child pornography and solicit minors for sex,” Attorney General Raul Torrez said in a statement Wednesday.

The civil lawsuit filed late Tuesday against Meta Platforms Inc. in state court also names its CEO, Mark Zuckerberg, as a defendant.

In addition, the suit claims Meta “harms children and teenagers through the addictive design of its platform, degrading users’ mental health, their sense of self-worth and their physical safety,” Torrez’s office said in a statement.

Those claims echo others in a lawsuit filed in late October by the attorneys general of 33 states, including California and New York, against Meta that alleges Instagram and Facebook include features deliberately designed to hook children, contributing to the youth mental health crisis and leading to depression, anxiety and eating disorders. New Mexico was not a party to that lawsuit.

Investigators in New Mexico created decoy accounts of children 14 years and younger that Torrez’s office said were served sexually explicit images even when the child expressed no interest in them. State prosecutors claim that Meta let dozens of adults find, contact and encourage children to provide sexually explicit and pornographic images.

The accounts also received recommendations to join unmoderated Facebook groups devoted to facilitating commercial sex, investigators said, adding that Meta also let its users find, share and sell “an enormous volume of child pornography.”

“Mr. Zuckerberg and other Meta executives are aware of the serious harm their products can pose to young users, and yet they have failed to make sufficient changes to their platforms that would prevent the sexual exploitation of children,” Torrez said, accusing Meta’s executives of prioritizing “engagement and ad revenue over the safety of the most vulnerable members of our society.”

Meta, based in Menlo Park, California, did not directly respond to the New Mexico lawsuit’s allegations, but said it works hard to protect young users with a serious commitment of resources.

“We use sophisticated technology, hire child safety experts, report content to the National Center for Missing and Exploited Children, and share information and tools with other companies and law enforcement, including state attorneys general, to help root out predators,” the company said. “In one month alone, we disabled more than half a million accounts for violating our child safety policies.”

Company spokesman Andy Stone pointed to a company report detailing the millions of tips Facebook and Instagram sent to the National Center in the third quarter of 2023 — including 48,000 involving inappropriate interactions that could include an adult soliciting child sexual abuse material directly from a minor or attempting to meet with one in person.

Critics, including former employees, have long complained that Meta’s largely automated content moderation systems are ill-equipped to identify and adequately eliminate abusive behavior on its platforms.
…

Spotify says it is planning to lay off 17% of its global workforce, amounting to around 1,500 employees, following layoffs earlier this year of 600 people in January and an additional 200 in June.

The music streaming giant is continuing its effort to cut costs and work toward becoming profitable, said Spotify CEO Daniel Ek in a prepared statement.

“By most metrics, we were more productive but less efficient,” he said. “We need to be both.”

The layoffs come following a rare quarterly net profit of about $70.3 million in October. The company has never seen a full year net profit.

“I realize that for many, a reduction of this size will feel surprisingly large given the recent positive earnings report and our performance,” Ek said. “We debated making smaller reductions throughout 2024 and 2025. Yet, considering the gap between our financial goal … and our current operational costs, I decided that a substantial action to right size our costs was the best option to accomplish our objectives.”

With the new layoffs, the company now expects to see a fourth quarter loss between $100 million to $117 million after previously anticipating a $40 million profit.

A majority of the charges will go toward severance for laid off employees, who will get about five months’ pay, vacation pay and health care coverage for the severance period.

Spotify did not clearly state when the layoffs would become financially beneficial but said that they would “generate meaningful operating efficiencies going forward.”

Spotify is following many companies in the tech industry trying to cut costs after growth in the industry slowed following a surge during the COVID pandemic.

Tech giants including Meta, Microsoft, Amazon and Google parent company, Alphabet, all have plans to cut 10,000 or more people this year.

Spotify began informing affected employees on Monday.

Some information in this report came from Reuters, The Associated Press and Agence France-Presse.
…

In early November, Nigeria launched its first interconnected solar hybrid mini grid in Nasarawa State to make electricity more reliable, renewable and accessible. Gibson Emeka has this story from Abuja, Nigeria, Grace Oyenubi narrates.
…

A small western Pennsylvania water authority was just one of many organizations breached in the United States by Iran-affiliated hackers who targeted a specific industrial control device because it is Israeli-made, U.S. and Israeli authorities say.

“The victims span multiple U.S. states,” the FBI, the Environmental Protection Agency, the Cybersecurity and Infrastructure Security Agency, known as CISA, as well as Israel’s National Cyber Directorate said in an advisory emailed to The Associated Press late Friday.

They did not say how many organizations were hacked or otherwise describe them.

Matthew Mottes, the chairman of the Municipal Water Authority of Aliquippa, which discovered it had been hacked on Nov. 25, said Thursday that federal officials had told him the same group also breached four other utilities and an aquarium.

Cybersecurity experts say that while there is no evidence of Iranian involvement in the Oct. 7 attack into Israel by Hamas that triggered the war in Gaza, they expected state-backed Iranian hackers and pro-Palestinian hacktivists to step up cyberattacks on Israeli and its allies in its aftermath. And that has happened.

The multiagency advisory explained what CISA had not when it confirmed the Pennsylvania hack Wednesday — that other industries outside water and water-treatment facilities use the same equipment — Vision Series programmable logic controllers made by Unitronics — and were also potentially vulnerable.

Those industries include “energy, food and beverage manufacturing and healthcare,” the advisory says. The devices regulate processes including pressure, temperature and fluid flow.

The Aliquippa hack promoted workers to temporarily halt pumping in a remote station that regulates water pressure for two nearby towns, leading crews to switch to manual operation. The hackers left a digital calling card on the compromised device saying all Israeli-made equipment is “a legal target.”

The multiagency advisory said it was not known if the hackers had tried to penetrate deeper into breached networks.

The advisory says the hackers, who call themselves “Cyber Av3ngers,” are affiliated with Iran’s Islamic Revolutionary Guards Corps, which the U.S. designated as a foreign terrorist organization in 2019.

The group targeted the Unitronics devices at least since Nov. 22, it said.

An online search Saturday with the Shodan service identified more than 200 such internet-connected devices in the U.S. and more than 1,700 globally.

The advisory notes that Unitronics devices ship with a default password, a practice experts discourage as it makes them more vulnerable to hacking. Best practices call for devices to require a unique password to be created out of the box. It says the hackers likely accessed affected devices by “exploiting cybersecurity weaknesses, including poor password security and exposure to the internet.”

In response to the Aliquippa hack, three Pennsylvania congressmen asked the U.S. Justice Department in a letter to investigate. Americans must know their drinking water and other basic infrastructure is safe from “nation-state adversaries and terrorist organizations,” U.S. Sens. John Fetterman and Bob Casey and U.S. Rep. Chris Deluzio said.

Cyber Av3ngers claimed in an Oct. 30 social media post to have hacked 10 water treatment stations in Israel, though it is not clear if they shut down any equipment.

Unitronics has not responded to the AP queries about the hacks.

The attack came less than a month after a federal appeals court decision prompted the EPA to rescind a rule that would have obliged U.S public water systems to include cybersecurity testing in their regular federally mandated audits. The rollback was triggered by a federal appeals court decision in a case brought by Missouri, Arkansas and Iowa, and joined by a water utility trade group.

The Biden administration has been trying to shore up cybersecurity of critical infrastructure — more than 80% of which is privately owned — and has imposed regulations on sectors including electric utilities, gas pipelines and nuclear facilities. But many experts complain that too many vital industries are permitted to self-regulate.
…

The U.S. proposed new guidelines Friday spelling out which electric vehicles will be eligible for tax credits, ruling out those that contain batteries or minerals sourced from China and other nations that have fallen out of favor with the U.S.

The restrictions dictate which clean energy vehicles will qualify for a subsidy of up to $7,500 under President Joe Biden’s Inflation Reduction Act, a federal law promoting sustainable, domestic energy production.

Only about 20 out of the more than 100 electric vehicles on the U.S. market qualify for a tax credit as it is. That number may be further reduced when this regulation goes into effect.

If a clean energy battery went through an assembly line owned by any “foreign entity of concern,” the car it will go into would be immediately disqualified from earning its owner any tax breaks from the U.S. government, starting in 2024.

The new rules target firms incorporated or headquartered in China, Russia, North Korea and Iran, among others, as well as companies where 25% or more of the equity interest or board seats are controlled by those countries.

From 2025 onward, electric vehicles made with critical minerals, such as lithium, nickel and cobalt, mined or processed by any “foreign entity of concern” will also be ineligible for subsidies.

The rules will be open to public feedback from automotive leaders for several weeks and are subject to change depending on industry recommendations.

Some information for this report came from Agence France-Presse. 
…

Developers at the University of Maryland are using a holographic camera to capture people’s movements in three dimensions for what could be high-impact training, education and entertainment. It is technology with the power to transform how we learn and entertain ourselves. VOA’s Julie Taboh has more. VOA footage by Adam Greenbaum.
…

US lawmakers warned Thursday that the Chinese Communist Party’s efforts to spread disinformation will only increase ahead of the 2024 elections in the United States. As VOA’s Congressional Correspondent Katherine Gypson reports, social media giant TikTok is lawmakers’ top concern.
Camera: Saqib Ui Islam
…