pause

Anthropic is proposing that the world’s top artificial intelligence companies come up with a coordinated way to pause development of advanced AI systems, warning that the technology is improving so quickly that there’s a risk humans would lose control.

The company behind the Claude chatbot said in a blog post on Thursday that, as cutting-edge AI gets increasingly faster at carrying out tasks, “it would be good for the world to have the option to slow or temporarily pause” its development.

Anthropic said its internal research institute plans to explore the issue in collaboration with others and “take actions” to help build the systems for a credible slowdown or pause, without being more specific.

Anthropic rival OpenAI argued for a different approach in a report published on Wednesday, saying that “democratic governments — not private companies acting alone — must ultimately determine the rules, safeguards, and accountability mechanisms”.

“Our view is that decisions about the pace of AI innovation should not be left to any one lab, company, or special interest group,” it said.

AI models are getting faster, with rapid increases in how quickly they can carry out software tasks like coding on their own, Anthropic said in its post. Based on current trends and given enough computing power, an AI system could be able to design and develop its own successor, in what is known as “recursive self-improvement”.

Self-building AI would be a major technological milestone that would bring benefits in science, healthcare and other areas, Anthropic said, but it “also might increase the risks of humans losing control over AI systems”.

Some tech industry figures have long warned of such a scenario.

Anthropic’s post comes after a different warning this week from a team of researchers at the University of Toronto who showed how AI tools could be used to create a new kind of AI “worm” that adapts its hacking strategy as it spreads from device to device and takes over a vast computing network.

“I think it’s really important that people understand that it’s not just the biggest, most powerful language models that pose the security concerns,” lead researcher Nicolas Papernot said in an interview.

The authors of the Anthropic post, company cofounder Jack Clark and Marina Favaro, head of its research institute, said the pause would be used to enable “societal structures and alignment research” to keep up with AI advances. Alignment is industry shorthand for making sure the technology matches human values and intentions.

The proposed coordination would let advanced AI labs verify that global rivals have actually stopped or slowed their work, “and that a bad actor could not use the auspices of a coordinated slowdown to jump ahead in secret”.

The company said a coordinated global mechanism is needed because, without it, a slowdown in AI development could let the “least cautious” players catch up and add to pressure on companies and governments as they make tough choices about AI safety.

Fears that advanced AI systems may get out of human control and cause societal harm have risen as the technology becomes increasingly capable. Anthropic’s own Mythos model sent shockwaves through industries, including banking and software, earlier this year with its ability to find vulnerabilities in existing code.

But regulation has been slow, especially in the US, where most leading AI labs are based. A Trump administration executive order earlier this week put the onus on the labs themselves, asking them to voluntarily submit their most capable models for government cybersecurity testing before public release.

Safety focus

AI researchers have also urged a pause before, but have had little success. Elon Musk, who owns AI lab xAI, was among the backers of a 2023 push by the non-profit Future of Life Institute to halt AI development for six months to allow time for safety guardrails.

Anthropic has long positioned itself as a safety-focused AI lab. Earlier this year, it refused to let the US military use its models for domestic surveillance and fully autonomous weapons, prompting backlash from the government, which put it on a national security blacklist, set to take effect later in 2026.

Anthropic’s post comes as the company and ChatGPT-maker OpenAI race to sell shares on the stock market, in an IPO that could value Anthropic at nearly a trillion dollars.

Papernot notified Canadian cybersecurity authorities prior to releasing his report, which shows how researchers developed the worm in a laboratory by using an “open-source” AI tool that is easy for software developers to cheaply access and modify.

“In the past, cyber attackers would focus on targets that are very high value,” he said. “Banking systems, hospitals, electricity grids, water treatment systems, schools.”

Papernot agreed that there should be more collaboration between companies, government agencies and academic researchers to develop countermeasures as AI-powered hacking tools supercharge the search for computer vulnerabilities.

“That old laptop you have in your basement that you don’t check on regularly doesn’t seem like a very high-value target, but it can be used as a launch pad to attack these higher-value targets,” he said. “Anything connected to the internet is now at risk because of how low the cost has become to mount these cyberattacks.”

May 2 (UPI) — A company that makes the abortion drug mifepristone on Saturday asked the U.S. Supreme Court to immediately pause a ruling that prevents doctors from prescribing it during telehealth visits.

Late Friday, a three judge panel on the 5th Circuit Court of Appeals unanimously ruled in favor of the state of Louisiana in a case asking the court to block doctors from prescribing the drug in telehealth visits.

Louisiana in the last four years has moved to prevent women in the state from obtaining abortion care legislators there were among the first to ban abortion after the repeal of Roe v. Wade, and later blocked doctors from prescribing the medical abortion pill in virtual telehealth visits.

The company, which is not the only drugmaker planning to file an appeal, said that patients will be stuck in limbo because of the lack of clarity it leaves for legal use of the drug, NBC News and Politico reported.

Roughly half of all abortions in the United States are performed using medications.

“Danco has been free to rely on procedures set by the FDA to distribute its product,” lawyers for the company said in a filing with the court.

“The Fifth Circuit’s decision immediately ends that,” the lawyers said. “A stay should issue to prevent the disruption and confusion that will result if the decision below were to remain operative.”

In addition to Danco, Politico reported that GenBioPro, which also manufactures the drug, has indicated that it will also file an appeal with the court.

Mifepristone was approved by the U.S. Food and Drug Administration in 2000 for medical termination of pregnancy and, until the COVID-19 pandemic, could only be prescribed during in-person appointments.

Early in the pandemic and the country locked down in an effort to stem the spread of the virus, doctors sued the FDA to allow them to prescribe mifepristone during telehealth visits.

The FDA temporarily changed the rule, but in 2023 adopted it permanently as some states started to restrict access to abortion and abortion services after the Supreme Court struck down Roe v. Wade.

Pharmaceutical companies and patient advocates warned that the restriction circumvents the FDA’s regulatory authority, which is based on evidence and data, and that it may offer a path for people to challenge other medications based on personal interest or opinion.

In the case of Danco, it also immediately filed the appeal because it is the only product it makes and “without a valid legal framework for distributing that product, Danco will lose its only source of revenue and may be unable to continue operating.”

The U.S. Navy says it is still aiming to see the AGM-88G Advanced Anti-Radiation Guided Missile-Extended Range (AARGM-ER) enter operational service this year. This is despite the announcement of a planned “strategic pause” in purchases of the missiles in the 2027 Fiscal Year. AARGM-ER is set to give Navy carrier air wings a critical boost in their ability to neutralize ever-more capable hostile integrated defense networks.

AARGM-ER has been in the works since the late 2010s. Northrop Grumman is the current prime contractor, through its previous acquisition of Orbital ATK. The Navy has ordered dozens of the missiles already. Hence, it was very surprising when the service’s latest proposed budget for the 2027 Fiscal Year, rolled out in full last week, included no request for funding to buy more AGM-88Gs due to the aforementioned pause. All of this, coupled with previous delays and technical issues encountered in testing, had prompted new questions about the future of the program.

“U.S. procurements for the AARGM-ER program are planned to resume once the system has successfully completed all necessary testing and software updates. Our immediate priority is ensuring the weapon passes these rigorous testing milestones to achieve Initial Operational Capability (IOC) in September 2026,” a Navy spokesperson told TWZ. “After validating the software and testing, the plan would be to ramp up production to clear a backlog of over 150 missiles, with U.S. procurements officially restarting in FY28 [Fiscal year 2028]. In the interim, FY27 production will be allocated to Foreign Military Sales to fulfill our commitments to five signed international cases.”

The spokesperson did not name the foreign customers in question. However, Italy is a full partner in the development of the AGM-88G. The U.S. government has also previously approved sales of the missiles to Australia, Finland, and the Netherlands. Norway has publicly announced its intention to purchase AARGM-ERs, as well. The U.S. Air Force is also set to acquire these missiles. We will come back to this later on.

The AGM-88 family, also known as the High-speed Anti-Radiation Missile (HARM), traces its roots back to the 1970s. The AARGM-ER is a major redesign of the preceding AARGM variant, also designated the AGM-88E. The AGM-88G features a completely redesigned body optimized for high speed and range, as well as a new, more powerful rocket motor and control actuation system.

Inside, the AGM-88G reuses the guidance and control systems from the AGM-88E. By extension, this means the AGM-88G retains the same multi-mode guidance capability of its predecessor, which includes a GPS-assisted inertial navigation system and a millimeter-wave radar seeker. The AARGM-ER’s primary target set is hostile emitters, especially air defense radars, but the guidance package is designed to allow it to find its mark even if they shut down and stop sending out signals to home in on. The AGM-88E also has a more general, secondary ability to strike targets on land or at sea, including by just being directed to hit a specified set of coordinates.

AARGM F-18

The Navy sees the AGM-88G entering service first integrated with its F/A-18E/F Super Hornet fighters and EA-18G Growler electronic warfare jets, both of which can already employ the AGM-88E. AARGM-ER’s boosts in speed and range are seen as critical to ensuring the survivability and effectiveness of those non-stealthy aircraft in the face of an ever more capable air defense threat ecosystem.

AARGM-ER is also sized to allow for internal carriage on F-35A and C variants. There are plans to eventually integrate it for external carriage on all three F-35 variants, as well as legacy F/A-18C/D Hornets, as well.

As noted, the development of the AARGM-ER has had to contend with technical issues and delays over the years. Originally, the goal was to reach IOC on F/A-18E/F and EA-18G in Fiscal Year 2023.

“The AARGM-ER experienced significant delays as a result of rocket motor, structural, and software problems discovered during testing,” the Government Accountability Office (GAO) said in a report published in June 2025. “Contracting officials noted that the program worked with the prime contractor to investigate the root causes of the identified deficiencies and implement corrective actions, including changes in the production process.”

“The program is still experiencing production delays as well. Since our last assessment, program officials stated that testing issues, supply chain challenges, and construction delays for a new production facility slowed completion of the first two production contracts by 1 year,” GAO’s report added. “We have found that starting production before demonstrating a system will work as intended – which the Navy did – increases the risk of discovering deficiencies that require costly, time-intensive rework.”

“In FY25 [Fiscal Year 2025], the [AARGM-ER] program attempted three IT [integrated test] weapon employment tests using F/A-18F aircraft against a threat-representative integrated air defense land target at the China Lake Range in California,” according to a separate report from the Pentagon’s Office of the Director of Test and Evaluation (DOT&E), released in March of this year. “AARGM-ER successfully completed one of the three weapon events but exhibited performance discrepancies during the other two, to include one event during which range safety terminated the weapon after release. No further weapons employment testing was accomplished in FY25 pending implementation of updates required to address the problems that were identified.”

DOT&E warned in that report that the IOC schedule for AGM-88G could slip further to the first quarter of Fiscal Year 2027, which starts on October 1 of this year.

Earlier this year, the Navy somewhat urgently put out a contracting notice saying it was exploring options for a new long-range anti-radiation missile. The stated requirements for this Advanced Emission Suppression Missile (AESM) were very much in line with how the AARGM-ER has been discussed in the past, with one notable exception: a new demand for the ability to engage targets in the air, as well as on the surface. You can read more about why that is significant here. With the Navy confirming that it is still pushing ahead on AARGM-ER, it remains unclear how exactly the service sees ASEM fitting into its broader plans. There does not appear to be any explicit mention of ASEM in the Navy’s latest budget request.

As noted, the U.S. Air Force is also in line to acquire AGM-88Gs. An AARGM-ER subvariant with “improved warhead/fuze” is set to serve as a bridge to the Stand-in Attack Weapon (SiAW), as well. Reportedly now designated the AGM-88J, SiAW is a derivative of the AARGM-ER being developed to provide a broader high-speed strike capability. The Air Force expects to primarily employ SiAW against time-sensitive and/or high-value assets on the ground, especially ballistic and cruise missile launchers, air and missile defense nodes, electronic warfare systems, and even anti-satellite weapons.

Despite the Navy’s “strategic pause” with AARGM-ER, the Air Force is asking for more funds to purchase additional SiAWs in Fiscal Year 2027. The Air Force has said in the past that it has been targeting 2026 for reaching IOC with SiAW on the F-35A. SiAW flight testing to date, at least that has been disclosed, has involved carriage by F-16 fighters, and it is possible the missile could be integrated operationally onto that aircraft and others, as well. As an aside, Northrop Grumman has also been pitching a ground-launched member of the AARGM-ER/SiAW family, called the Advanced Reactive Strike Missile (AReS).

As mentioned, the Navy has made clear that procurement of AARGM-ERs for foreign customers through the FMS program is also continuing.

Time will tell whether or not the Navy can meet its IOC target for AARGM-ER by September, or the timeline slips into the next fiscal year. Still, the service looks to remain committed to the program, at least for the time being, regardless of its intent to put a year-long pause on buying more AGM-88G.

Contact the author: joe@twz.com