Move comes after US photojournalist and group of European diplomats and Palestinians were assaulted in separate attacks

Israeli police have detained 10 people in a rare raid against settlers, after a weekend of extreme violence in the occupied West Bank including an attack on US photojournalist Heidi Levine, which she described as an “attempted lynching”.

She was knocked down, kicked and beaten by a group of masked Israeli men as she covered the olive harvest in a Palestinian village on Friday. The militants broke four vertebrae and a rib, ripped open her finger so it required surgery and left her with lacerations across her body.

• WalletHub released a report ranking the safest cities in the US.
• The ranking analyzed 182 US cities based on 41 safety indicators across three categories.
• Juneau, Alaska, was ranked the safest city in the US, while New Orleans was ranked the least safe.

American cities face a range of safety concerns, and the risks can vary significantly depending on location.

In some cities, natural disasters such as hurricanes, tornadoes, and floods pose the greatest threats to locals. In others, safety concerns may instead center on financial security or the overall safety of homes and communities.

WalletHub compared a total of 182 cities — the 150 largest cities in the US, plus at least two of the most populous cities in every state — to determine the safest and least-safe cities in the US.

WalletHub examined 41 safety indicators across three categories: home and community safety, natural-disaster risk, and financial safety.

To evaluate home and community safety, WalletHub looked at factors including crime rates, traffic and pedestrian fatalities, and the number of emergency responders per capita.

The natural-disaster risk ranking was determined by cities' risk-index scores for natural disasters such as earthquakes, tornadoes, and wildfires.

The financial safety ranking considered factors such as poverty and unemployment rates, as well as identity theft complaints per capita.

To conduct its analysis, WalletHub looked at data from sources including the US Census Bureau, the Bureau of Labor Statistics, the Federal Emergency Management Agency, and the Federal Bureau of Investigation.

Here are the top 10 safest and least-safe cities in the US.

1. Juneau, Alaska

Out of the 182 cities WalletHub examined, Juneau was ranked the safest city in the US. It ranked 46th in home and community safety, fourth in financial safety, and first in natural-disaster risk.

Juneau featured the lowest murder rate and lowest pedestrian fatality rate, and there have been no recorded terrorist attacks or terrorist-related incidents in the past decade. The risk of natural disasters in Juneau is particularly low, with the second-lowest risk of flooding and tornadoes.

Financially, Juneau has the 21st-lowest percentage of residents below the poverty line.

2. Warwick, Rhode Island

As the second-safest city in the US, Warwick's home and community safety was ranked ninth, its natural-disaster risk was ranked fifth, and its financial safety was ranked 46th.

Warwick had the fourth-lowest number of aggravated assaults per capita and the sixth-lowest number of thefts per capita.

3. Overland Park, Kansas

Overland Park rounded out the top three safest cities in the US, with its home and community safety ranking fourth and its financial safety ranking seventh.

The Kansas City suburb placed 138th in WalletHub's natural-disaster-risk category. The greater Kansas City area had received 88 tornado warnings by June 2026, ABC affiliate KMBC reported.

However, Overland Park had the 19th-lowest pedestrian fatality rate and the 15th-lowest traffic fatality rate among the 182 cities WalletHub analyzed. It also had the 14th-lowest unemployment rate in the country and the third-lowest share of residents who are living in poverty.

4. Burlington, Vermont

Burlington is relatively safe from natural disasters, with its natural-disaster risk ranked third overall. Its financial safety was ranked 24th, and its home and community safety was ranked 43rd.

5. South Burlington, Vermont

South Burlington, Vermont, tied with Burlington for third place in natural-disaster risk. The two cities are only miles apart.

South Burlington's financial safety ranked higher, in sixth place. Its home and community safety ranked lower than Burlington's, at 51st.

6. Virginia Beach, Virginia

WalletHub ranked Virginia Beach 22nd in home and community safety, 10th in natural-disaster risk, and 55th in financial safety.

7. Cedar Rapids, Iowa

In Cedar Rapids, the city's home and community safety ranked 10th, its natural-disaster risk ranked 23rd, and its financial safety ranked 42nd.

8. Casper, Wyoming

Casper's financial safety ranking placed it in the top 10, in ninth place. Its natural-disaster risk ranked 19th, and its home and community safety ranked 32nd.

9. Boise, Idaho

Boise was ranked No. 33rd in home and community safety, 41st in natural-disaster risk, and fifth in financial safety.

10. Yonkers, New York

WalletHub's ranking named Yonkers as the 10th-safest city in the US. Out of 182 cities, it ranked third in home and community safety, 59th in natural-disaster risk, and 108th in financial safety.

Keep reading for the bottom 10 cities, which are ranked the least safe in the US.

173. St. Louis, Missouri

Out of the 182 cities WalletHub examined, St. Louis ranked 177th in home and community safety and 146th in financial safety.

The Missouri city ranked relatively well in its natural-disaster risk ranking, however, coming in at 34th.

174. Cleveland, Ohio

Cleveland ranked well in WalletHub's natural-disaster risk rating, coming in at 74th.

In the financial safety category, the city ranked 153rd. The lowest category it was ranked in was the home and community category, ranking 175th.

175. Houston, Texas

Houston's financial safety ranking placed the city 163rd. Its natural-disaster risk ranked 170th, and its home and community safety ranked 171st.

176. Jacksonville, Florida

Jacksonville was ranked 170th in home and community safety, 177th in natural-disaster risk, and 166th in financial safety.

177. Detroit, Michigan

WalletHub ranked Detroit 173rd in home and community safety and 67th in natural-disaster risk.

It was ranked the least safe city in the US for financial safety, coming in at 182nd.

178. Philadelphia, Pennsylvania

In Philadelphia, the city's home and community safety ranked 176th, its natural-disaster risk ranked 137th, and its financial safety ranked 159th.

179. Memphis, Tennessee

WalletHub ranked Memphis as the second-worst city for financial safety, coming in at 181st, behind Detroit.

The city ranked 179th in home and community safety, and jumped to 60th in the natural-disaster risk category.

180. Baltimore, Maryland

Baltimore was ranked 180th in home and community safety. It ranked 174th in financial safety, and performed better in the natural-disaster risk category, placing 106th.

181. Baton Rouge, Louisiana

Ranked the second-least-safe city in the US, Baton Rouge came in 181st for home and community safety, 130th for natural-disaster risk, and 144th for financial safety.

182. New Orleans, Louisiana

Out of the 182 cities WalletHub examined, New Orleans is considered the least safe city in the US, largely due to its home and community safety, where it ranked the least safe, and its financial safety, where it ranked 172nd.

As a hurricane-prone city, New Orleans ranked better in the natural-disaster risk category, however, coming in at 104th.

• Anthropic is looking to shape AI regulation by pitching an AI safety bill of its own.
• POLITICO obtained a confidential draft of the proposed legislation.
• The bill would let the government block an AI model only if it poses a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people.

Representatives for the artificial intelligence giant Anthropic have quietly pitched Capitol Hill lawmakers on a draft bill to regulate frontier AI safety and security, according to three people with knowledge of the efforts who were granted anonymity to discuss private deliberations.

The confidential draft, obtained by POLITICO and not previously reported, marks a notable effort by the frontier AI lab to shape the rules governing its own technology as Congress weighs how to respond to the fast-moving industry.

"This bill represents the proposals outlined in Anthropic's Advanced AI Framework, which we have been advocating for with federal and state lawmakers since June," an Anthropic spokesperson said. "These are sensible policies we believe Congress should act on."

The bill would let the government block a model's release or use only if it poses a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people. The risk must stem from the creation or release of a biological weapon, the model evading its developer's controls or a cyberattack. The language mirrors that found in California's AI laws including SB 53, which was signed into law last year.

Leading AI labs disclosed a series of incidents this summer, in which AI agents hacked or probed websites and platforms. None reported so far appear to have caused widespread deaths or over $1 billion in damages.

Anthropic has yet to publicly back particular legislative efforts in Congress, but they have backed some state AI safety laws, including two in California. Their head of U.S. state and local government relations previously told POLITICO that they hope states will continue to pass ever-stronger AI safety laws that raise the standard nationwide.

Under the bill, independent evaluators, which would register with the Commerce Department, would be tasked with periodically evaluating the lab and its frontier models. The evaluators may be paid for their work by the lab itself so long as the payment is not contingent on the evaluations' results.

If the evaluators determine there is "imminent enumerated risk" for a company's frontier models, they must report it to Commerce within 72 hours of completing their assessment.

While the Commerce Department could bring a request to the judge to halt the release of a model after determining that it poses a risk, the language explicitly prohibits any other U.S. "department, agency, or officer" from suspending the model's development, deployment or internal use. A federal judge for the District Court for the District of Columbia could order the suspension of a model.

Over the summer, Anthropic was subject to sweeping export controls that shut down two of its most advanced models. The decision was made in a whirlwind series of phone calls between administration officials and the company's executives, worrying some in the tech industry that access to their technology may be restricted on the whims of government officials.

Anthropic's bill would also preempt some state AI laws related to public disclosure of risk, third-party evaluation and reporting of critical incidents. Preemption has been a political flashpoint on Capitol Hill, with Democrats rejecting some proposals they feared would neuter regulations passed by the likes of California, Illinois and New York.

Many tech industry groups have argued that a state-by-state patchwork of laws is difficult to comply with compared to a unified national standard.

The legislation would also require frontier labs to create safety frameworks, publish transparency reports for new frontier models, implement protections for whistleblowers and report critical safety incidents to Commerce.

Capitol Hill has seen a flurry of activity on AI safety legislation this summer. In the House, Reps. Lori Trahan (D-Mass.) and Jay Obernolte (R-Calif.) spearheaded the FRONTIER Act, which would give the government the ability to stop the deployment of risky models and preempt some state AI laws.

In the Senate, Majority Leader John Thune and Sens. Amy Klobuchar (D-Minn.) and Ted Cruz (R-Texas) have been negotiating an AI safety bill intended to mitigate catastrophic risks. But it has faced headwinds from some safety-minded groups and Sen. Maria Cantwell (D-Wash.), the top Democrat on Cruz's Commerce Committee, who want the government to have a larger role in reviewing AI models.

Other progressives, including Rep. Greg Casar (D-Texas) and Sen. Bernie Sanders (I-Vt.), have gone well beyond their colleagues in calling for a pause in advanced AI development and a ban on superintelligence, a type of theoretical AI that far surpasses humans' cognitive capabilities.

This story originally appeared on POLITICO and is courtesy of the Axel Springer Global Reporters Network, which harnesses the resources of the company's newsrooms to publish ambitious scoops, investigations, interviews, opinion pieces, and analysis. It allows journalists — including those from POLITICO, Business Insider, WELT, BILD, Onet, and Fakt — to collaborate on major stories for an international audience of hundreds of millions across platforms.

Exclusive: Using test accounts acting as 13-year-olds, Tech Transparency Project downloaded 135 apps with features to generate nonconsensual nude imagery

In December 2025, Morgan Dorfman received distressing news: a student at her daughter’s Pennsylvania high school was making AI-generated pornographic material using the daughter’s image.

He was using an artificial intelligence image-editing app called Movely to create sexual videos from pictures of her daughter and five other girls.

Mike Huckabee calls for arrest and prosecution of those responsible as Heidi Levine, 63, recovers in hospital in Israel

The US ambassador to Israel, Mike Huckabee, has denounced the settler attack that left the Washington Post photojournalist Heidi Levine badly beaten and needing surgery as a “despicable act of terror” and called for the arrest and prosecution of those responsible.

Levine, 63, an award-winning journalist who lives in Israel, was set upon last Friday by masked Israeli settlers who used clubs and other weapons to attack Palestinian people attempting to harvest their olives in occupied West Bank towns.

A Russian drone tore a hole in the shield over Chornobyl’s dead reactor. Now the G7 is paying to patch it.

The Group of Seven nations plans to allocate 260 million euros ($292 million) to repair the New Safe Confinement at the Chornobyl Nuclear Power Plant, First Deputy Prime Minister and Energy Minister Denys Shmyhal says.

The confinement, a 35-story steel arch sealing the reactor that exploded in the 1986 disaster, was damaged by a Russian drone strike in February 2025. Full restoration will cost at least 500 million euros, the European Bank for Reconstruction and Development (EBRD) estimates, so the G7 pledge covers just over half, leaving about 240 million euros still to raise.

Shmyhal calls the decision an important step toward restoring nuclear safety at the plant, which the Russian strike had compromised. The money will go to priority work, including resealing the structure and restoring its ventilation system.

“Thank you to France and personally to Minister Roland Lescure for leadership in mobilizing international support,” Shmyhal says, crediting France and the EBRD with co-chairing the effort.

The night a drone pierced the sarcophagus

The damage dates to 14 February 2025. A Russian drone carrying a high-explosive warhead punched a hole through the arch’s roof and ignited a fire in the cladding, breaching the structure’s seal and ventilation.

The attack caused a 15 m² breach in the inner and outer cladding and damaged around 200 m² of panels.

"It was a targeted Russian attack on the Chornobyl station with a drone that is an analog of the Iranian Shahed," Ukrainian President Zelenskyy said at that time.

A bill only half-covered, and a 2030 deadline

The G7 money closes only part of the gap. Partners must still raise roughly 240 million euros more, and Ukraine hopes to secure it at a donor conference in Paris co-chaired by Ukraine, France and the EBRD, Shmyhal says.

The clock matters as much as the cost. The work must be finished by 2030 to limit corrosion and protect the arch’s 100-year design life, the EBRD says, turning a repair bill into a race against the rust the strike set loose.

We went aboard a cruise ship navigating some of the world's most dangerous waters. In the Arctic, drifting icebergs, sea ice, sudden storms and fierce winds make sailing unpredictable. Yet every year, around 100,000 cruise passengers travel to Greenland to explore its remote coastline and experience one of the world's last great wildernesses.

On Friday, Gabriela Castillo Madrid, SpaceX’s Senior Manager of Global Government Affairs, is set to appear before the New York City Council’s Committee on Technology to answer questions about the safety of its artificial intelligence technology.

RELATED:

Artificial Intelligence Pioneers Call for International Body to Oversee Cybersecurity Standards

The hearing comes after City Council Speaker Julie Menin formally subpoenaed the company to participate in an Oct. 5 hearing on AI risks, which it did not attend.

After the initial refusal, the body formally notified the company of its intent to file a petition with the New York State Supreme Court to compel compliance, after which the firm requested a meeting and agreed to testify.

SpaceX’s appearance follows the Oct. 5 hearing, which included senior representatives from OpenAI, Anthropic, Meta and Google, as well as sworn testimony from prominent industry whistleblowers Jacob Coxon, Daniel Kokotajlo and Alex Turner, and British lawmaker Jess Asato.

“We held a multi-hour hearing with whistleblowers, the world’s largest technology companies and the nation’s leading AI experts. The conspicuous absence of SpaceXAI was unacceptable,” Menin said at the time.

Citizens deserve “to hear directly from the companies developing this powerful technology what safeguards they are implementing to protect people and how they will be held accountable when they fail,” she pointed out, warning that “self-regulation is not enough.”

During the previous hearing, Menin asked spokespeople for OpenAI, Anthropic, Meta and Google under oath to commit to not releasing a model if it failed an internal test or an evaluation by a third party outside the company.

teleSUR/ JF

Source: EFE

Settlers hurl rocks at Palestinians and beat photographer Heidi Levine as part of yearly attacks to disrupt harvest

Israeli settlers have attacked Palestinians across the occupied West Bank in a coordinated movement to disrupt the annual olive harvest, burning cars and trees and injuring Palestinians and a Washington Post photographer.

Masked settlers wielding clubs and other weapons attacked people attempting to harvest their olives in West Bank towns on Friday. Palestinian authorities said they recorded attacks in eight villages, and injuries among at least 20 people.

Thanks to a 2005 DUI conviction in Pennsylvania, Edward Williams permanently lost the constitutional right to keep and bear arms. But he did not realize that until a decade later, and his confusion on that point is understandable.

The federal law that stripped Williams of his Second Amendment rights is commonly described as applying to "felons," and Williams' offense was a misdemeanor. But although his sentence did not include any time behind bars, his crime theoretically could have sent him to prison for up to five years. That was enough to trigger 18 USC 922(g)(1), which covers anyone convicted of "a crime punishable by imprisonment for a term exceeding one year."

When a lawyer alerted Williams to that wrinkle, which made his continued ownership of firearms a federal felony, he got rid of his guns. But he did not think that requirement was fair, reasonable, or constitutional, and this week a federal appeals court agreed. The U.S. Court of Appeals for the 3rd Circuit recognized a historical tradition of disarming "physically dangerous individuals" whose possession of firearms would pose a threat to public safety. But it said Williams plainly does not fit into that category.

"The federal government spent nearly a decade trying to permanently disarm a peaceable man over a nonviolent misdemeanor—and it lost bigly," says Brandon Combs, president of the Firearms Policy Coalition, which supported Williams' lawsuit. "The government does not get to turn a nonviolent conviction into a lifetime ban on exercising a constitutionally protected right."

The 3rd Circuit's en banc decision in Williams v. Attorney General affirms a 2023 ruling by a federal judge who likewise concluded that disarming Williams was unconstitutional. It elaborates on the logic of Range v. Attorney General, a 2023 decision that restored the gun rights of a Pennsylvania man who had been convicted of food stamp fraud. In both cases, the 3rd Circuit held that applying Section 922(g)(1) to the plaintiffs was not "consistent with this Nation's historical tradition of firearm regulation"—the Second Amendment test established by the Supreme Court's 2022 decision in New York State Rifle & Pistol Association v. Bruen.

Writing for the majority in Williams' case, 3rd Circuit Judge Paul Matey acknowledges that the Supreme Court has described "longstanding prohibitions on the possession of firearms by felons" as "presumptively lawful." But "those longstanding customs restricted the violently dangerous (within which some felons would be included), not felons writ large," he says. "Felons were restricted because they posed a physical danger, rather than for their offense of conviction."

Matey argues that "focusing on physical dangerousness corrects the erroneous argument" that "only law-abiding and responsible citizens can exercise the rights protected by the Second Amendment"—a claim that the Supreme Court rejected in the 2024 case United States v. Rahimi. "We find no history that supports disarming irresponsible people, where 'responsible' is defined by the whim of the sovereign or the will of the majority," he says, alluding to the danger of allowing legislators to whittle away at the Second Amendment by criminalizing an ever-widening range of conduct.

Matey emphasizes "the novelty of forever restricting all felons from possessing firearms for lawful purposes." The original restriction, which Congress approved in 1938 as part of the Federal Firearms Act, applied only to violent crimes such as murder, manslaughter, rape, kidnapping, robbery, and assault with a deadly weapon. In 1961, Congress expanded the ban to cover nonviolent crimes punishable by more than a year in prison.

Williams' 2005 DUI conviction qualified for two reasons. He had a prior DUI conviction, stemming from a 2000 arrest in State College. And when he was arrested again in Philadelphia four years later, his blood alcohol concentration was a whopping 0.223 percent. That made his second offense a first-degree misdemeanor, which carries a maximum prison sentence of five years.

Since then, Williams has stayed out of trouble. "He has done nothing violent, let alone dangerous, for over twenty years, and has held steady employment at the same job for over three decades," Matey notes. Williams drinks only occasionally, and a psychological evaluation found he "has a normal personality, without psychopathology and without addiction or violent tendencies."

The Justice Department nevertheless argued that Williams' continuing disarmament is consistent with the Bruen test, which requires historical analogs that are "relevantly similar" in motivation and scope. In making that case, the government relied mainly on early laws aimed at "habitual drunkards"—the same analogy that the Supreme Court unanimously rejected last June in United States v. Hemani, which involved the federal ban on gun possession by "unlawful" drug users.

Although the Justice Department "acknowledges it may proscribe Williams's possession of firearms only if he 'presently poses a physical danger if armed,'" Matey says, it "offers no evidence to establish this as the aim of the habitual drunkard laws." Those laws, the Supreme Court noted in Hemani, "did not seek to protect the public from violence so much as to protect habitual drunkards from themselves and their families from financial devastation."

Nor did the habitual drunkard laws permanently disarm people, as Section 922(g)(1) does. "Most statutes capped the commitment of habitual drunkards to the state inebriate asylum, prison, or hard labor to a few months or years," Matey notes. "Others provided for an indefinite period of incarceration, but had no indication that upon release individuals continued to forfeit their arms. Rather, once someone was 'safe' to 'go at large,' they were restored to the status of a full rights-bearing citizen."

Contrary to the considerations that the Supreme Court deemed relevant in Hemani, Matey says, the Justice Department "fails to show that Williams threatens to 'misuse[] a gun while intoxicated,' or that he 'poses a danger to himself or others as a result' of
alcohol abuse. In fact, much like those nineteenth-century drunkards released from custody and restored their rights, the record shows that Williams poses no present physical danger from the use of a firearm, or anything else."

The 3rd Circuit is one of three federal appeals courts that have allowed people with nonviolent felony records to argue that Section 922(g)(1) is unconstitutional as applied to them. But such challenges have been blocked in six circuits, notwithstanding compelling historical evidence that the law reaches much farther than the Second Amendment allows. So far, the Supreme Court has declined to accept any of the many petitions asking it to resolve that circuit split.

The post The 3rd Circuit Upholds the Gun Rights of a Man Who Committed a Nonviolent Misdemeanor 2 Decades Ago appeared first on Reason Magazine.

A group of 119 members of the European Parliament is pressing the European Commission and national governments to ensure that a new EU ban on artificial-intelligence  ‘Nudifier’systems used to generate non-consensual sexual images can be enforced as soon as it takes effect.

The prohibition on so-called AI “nudifier” systems becomes applicable on 2 December 2026 following changes to the EU Artificial Intelligence Act approved earlier this year.

The legislation prohibits AI systems designed to create child sexual abuse material or realistic images, video or audio depicting an identifiable person’s intimate parts or sexually explicit activity without that person’s consent.

Providers cannot place systems intended for such use on the EU market. Systems capable of producing the material must have adequate safeguards intended to prevent it, while deployers are prohibited from deliberately using AI systems for that purpose.

The final agreement on the AI Act amendments gave companies until 2 December to bring their systems into compliance.

The latest intervention is intended to ensure that regulators are ready to enforce the prohibition rather than simply having the legislation on the statute book.

The letter supported by 119 MEPs was initiated by Green MEP Lena Schilling and Greens/EFA co-president Terry Reintke.

Its signatories include members from several political groups.

They want the Commission and member states to ensure that national authorities are designated to enforce the ban and that people targeted by non-consensual AI-generated sexual imagery know which authority to contact.

The letter also calls for stronger enforcement of the Digital Services Act against online platforms and search engines carrying or facilitating access to such material.

The initiative follows research cited by the organisers which identified 147 European politicians appearing on websites containing sexualised AI deepfakes.

Of those identified, 138 were women.

Schilling was among those affected and has used her own case to argue that the December prohibition needs an operational enforcement system from its first day.

The prohibition emerged during the EU’s broader revision of its AI legislation.

In March, the European Parliament adopted its negotiating position by 569 votes to 45, with 23 abstentions.

Parliament and Council subsequently reached agreement on the revised text, and the measure completed its formal legislative process in June.

EU Today reported at the time that the Council’s final approval combined delays to some high-risk AI requirements with new restrictions on non-consensual sexual imagery.

The issue is distinct from the wider transparency requirements applying to AI-generated content.

Under Article 50 of the AI Act, transparency obligations have applied since 2 August.

Providers of generative systems must in relevant cases make AI-generated or manipulated material machine-readable and detectable, while deployers face disclosure obligations covering deepfakes and some AI-generated material dealing with matters of public interest.

EU Today examined those changes when chatbot and deepfake disclosure rules entered into application in August.

There is, however, a transitional period for some AI systems that were already on the market before 2 August. Their machine-readable marking obligations take effect on 2 December.

That means the same date will bring both the new nudifier prohibition and the end of the transitional period for certain AI-content marking requirements.

Enforcement is divided between national regulators and EU institutions.

National market-surveillance authorities carry much of the responsibility for enforcing the AI Act. The Commission’s AI Office has direct enforcement responsibilities in more limited areas, including certain systems built on general-purpose AI models and systems integrated into very large online platforms or search engines.

The Commission’s guidance on Article 50 is intended to provide regulators and companies with a common interpretation of those transparency requirements.

The nudifier prohibition raises an additional problem because distribution of non-consensual intimate images also intersects with the Digital Services Act, national criminal and civil laws and EU legislation addressing violence against women.

The 119 MEPs are therefore seeking action beyond the AI Act itself.

Their letter asks for stricter use of the Digital Services Act against platforms and search engines, faster implementation of the EU directive combating violence against women and clearly identifiable national authorities able to respond to complaints.

The legal gap surrounding nudifier systems had previously been acknowledged by the Commission.

Its review of prohibited AI practices concluded that the original AI Act’s provisions on manipulation and exploitation of vulnerabilities did not adequately cover systems producing child sexual abuse material and non-consensual sexually explicit images.

That finding contributed to the explicit prohibition subsequently added during the Digital Omnibus negotiations.

The legislation is now settled. The immediate issue before the 2 December deadline is whether regulators across the 27 member states have the mechanisms in place to identify prohibited systems, investigate providers and respond when people report AI-generated intimate material produced without their consent.

Fear and anger are growing in South Africa after the bodies of nine women were discovered across the Ekurhuleni municipality, east of Johannesburg, over the past two months, with police investigating whether the deaths are connected, a 10th body was confirmed this week. Three suspects have been arrested, and police have acknowledged a possible connection between some of the killings. What is not in dispute is the scale of the crisis the discovery has exposed, and the outrage it has provoked, protests against femicide have been held in the country in the weeks since.

Four of the nine women have so far been publicly identified, Elizabeth “Tsontso” Moselakgomo, 38, a mother who vanished during her regular afternoon run and was found days later behind a hotel in Rhodesfield; Itumeleng Kekana, 32, missing since July and found in August with multiple stab wounds; Dineo Evelyn Motapane, 38, found in KwaThema; and Gracious Nkomo, 28, a lodge worker in Kempton Park missing for more than two weeks before she was found. 

Anger over the response

The deaths have triggered protests and renewed criticism of the country’s response to gender-based violence. Residents and women’s rights campaigners have questioned why women are repeatedly advised to change their behavior, avoid isolated areas, walking alone or running alone, rather than placing greater emphasis on preventing perpetrators from committing violence.

The National Union of Metalworkers of South Africa (NUMSA) has been among the sharpest voices responding to the killings. In a statement condemning the deaths, the union insisted the killings should not be treated as a “crime wave” or a mystery, describing them instead as femicide and, more precisely, a war being waged against women. NUMSA’s argument is that this is not an aberration but the visible edge of something structural. The union points to police figures showing at least 569 women murdered nationally between April and June 2026 alone, by the verified count, with an unverified estimate closer to 892, roughly six women killed every day. Against this, nine deaths in one metro over two months is, in the union’s words, closer to the ordinary condition of the country, made visible only because the fear of a serial killer has made it a story the wider public can imagine touching them.

The statement also pushes back on the instinct to treat this primarily as a street-level danger. Of thousands of rape cases registered by police in the same quarter, the majority occurred in a residence, not a public place. The killings that draw headlines, a body beside a highway, represent a fraction of a type of violence whose center of gravity is the home.

The economics underneath the violence

NUMSA also refuses to treat gender-based violence as an individual pathology alone. It traces a line from the colonial and apartheid-era migrant labor system, which depended on women’s unpaid domestic and reproductive labor to keep wages artificially low through to the present, where South African women remain disproportionately unemployed (an official rate near 37.5%, against 44.9% for men) and disproportionately head of household. Over four in ten nationally, and closer to half in rural areas. Close to two million employed women work informally, without pensions, medical aid, or written contracts. When a woman who is also a breadwinner is killed, the union argues, it is not only a life lost but a household’s income and caregiving capacity lost with her.

Gender-based violence and femicide were declared a national disaster in South Africa in November 2025. Nearly a year on, the killings in Ekurhuleni are evidence, in the union’s reading, that the declaration has not translated into protection.

What’s being demanded

NUMSA has rejected the idea of policing alone and the state’s response so far has leaned on a security “lockdown” of Kempton Park to answer a crisis this size. Its demands include full funding of the national plan on gender-based violence and femicide, functioning forensic and mortuary capacity so victims can be identified quickly, safe transport for workers on early and late shifts, and the implementation of the ILO Convention 190 on violence and harassment in the workplace, which South Africa has ratified but not fully enforced. Alongside state accountability, the union is calling for community-based structures, street and area committees, explicitly modeled on the civic organizing of the anti-apartheid period to accompany women, support survivors, and track cases through the courts.

Investigations continue, but the fear, concern and anger surrounding the killings remain. The deaths of the women in Ekurhuleni have already forced a public reckoning with the scale of violence against women in South Africa and the urgent need for stronger action to protect women and hold perpetrators accountable.

The European Parliament has backed calls for an EU-wide strategy to tackle worsening shortages of healthcare professionals, including an ambition to increase the health workforce by one million, as doctors, nurses, pharmacists and dentists urged the European Commission to turn the recommendations into concrete action.

The own-initiative report, led by MEPs Loucas Fourlas (EPP, Cyprus) and Ruggero Razza (ECR, Italy), warns that shortages of doctors, nurses, carers and other professionals are becoming structural, driven by ageing populations, rising demand for care and persistent burnout...