RobinsonLabs/ news

Trump tells Axios he no longer views Anthropic as a national security threat

The U.S. president softened his stance a week after ordering Anthropic to block foreign access to its most advanced models.

President Donald Trump told Axios on Friday that he "no longer" views Anthropic as a national security threat, reversing a position he acknowledged holding just days earlier. The pivot follows a dispute over the company's Fable 5 and Mythos 5 frontier models, which Anthropic globally disabled last week after a White House export-control directive demanded it block all foreign nationals from accessing the technology. Senior Anthropic technical staff met with administration officials earlier in the week to discuss the matter, and Trump told Axios that CEO Dario Amodei had responded "very quickly" and "responsibly" to the directive.

Trump did not rule out using emergency powers under the Defense Production Act against Anthropic, according to Axios, noting he "have[s] the power to use a lot of things" but was "not sure [he] has to do that." The G7 summit in France this week also brought Trump and Amodei together with other tech CEOs. An Anthropic spokesperson said the company is "grateful to the administration for their ongoing partnership in working to get this matter resolved as quickly as possible."

Norway imposes near-total ban on AI in elementary schools

The country moved to protect foundational learning skills after a broader decline in student test scores.

Norwegian Prime Minister Jonas Gahr Store announced on Friday that pupils aged 6 to 13 will be barred as a general rule from using generative AI in schools starting in late August. Students aged 14 to 16 will be allowed to use AI tools only under direct teacher supervision, while older students aged 17 to 19 will be taught to use AI "appropriately" to prepare them for further education and work. The government also announced plans to propose legislation that would promote the use of books in classrooms, reversing its earlier push toward digital tablets.

The policy comes after a broad decline in Norwegian education test scores, which the government has already sought to address by banning smartphones in schools in 2024. Store said at a press conference that using AI "increases the risk that young children skip important steps in their education," adding that "the most important thing in school is that our children learn to read, write and do mathematics."

Pew study finds only 16 percent of Americans think AI will have a positive impact

A new survey reveals deep skepticism about the technology's long-term effects across the population.

A new Pew Research study shows that only 16 percent of Americans believe AI's impact on society over the next 20 years will be positive, while roughly 40 percent say it will be negative. A vast majority of respondents (67 percent) do not believe the U.S. government will meaningfully regulate AI, and 59 percent say they do not trust companies to develop the technology safely. Nearly two-thirds of Americans also think AI is developing too quickly. Young Americans hold the most negative views: only 14 percent of those under 30 expect a positive impact.

Despite the widespread skepticism, AI adoption continues to grow. About 25 percent of Americans now use AI chatbots daily, and 44 percent of U.S. adults have used ChatGPT at least once — more than double the figure in 2023. Pew found that men still use AI tools more frequently and are more enthusiastic about them than women. Six in ten adults say they routinely read AI-generated internet summaries, though a large share of Americans — nearly half — still do not use AI in their daily lives, with older Americans least likely to engage with the technology.

German court rules Google directly liable for AI Overviews

A landmark ruling in Munich strips search-engine protections from Google's AI-generated content.

The Munich Regional Court ruled on Friday that Google is directly liable for false and defamatory claims generated by its AI Overview feature, rejecting the company's argument that it should be treated as a passive search engine. The court determined that AI Overviews constitute Google's own content because they rewrite, judge, and synthesize third-party results "in their own words and according to their own structure" rather than simply listing links. In the case at hand, the AI incorrectly linked two Munich publishing companies to scam operations and fraudulent schemes.

The court held that AI-generated statements have lower constitutional protection than human expression since they are "the result of an algorithm" rather than "an expression of an acquired conviction." Traditional search-engine liability frameworks and Digital Services Act host-protection exemptions do not apply, and Google was ordered to stop spreading the false claims and to bear 80 percent of the litigation costs. Google said it would appeal the ruling, which is not yet final, and stated that AI Overviews "are designed to reflect the information that exists on the web."

Salesforce acquires AI agent platform Fin for $3.6 billion

The deal bolsters Salesforce's Agentforce ecosystem as enterprise AI competition intensifies.

Salesforce announced on Monday that it will acquire AI customer-service platform Fin (formerly Intercom) for approximately $3.6 billion. Fin's flagship product is an autonomous AI agent powered by its proprietary Apex model, capable of resolving customer queries across chat, email, WhatsApp, SMS, phone, and Slack. The acquisition is intended to strengthen Salesforce's Agentforce platform, which allows enterprises to build and deploy custom AI agents for business workflows. The deal is expected to close in the fourth quarter of Salesforce's 2027 fiscal year.

Salesforce shares have shed more than a third of their value in 2026 as investors worry AI could disrupt traditional SaaS business models. CEO Marc Benioff told CNBC that the company has seen "record" transactions in its latest quarter and argued that AI is an opportunity rather than a threat. Fin CEO Eoghan McCabe said little would practically change under the deal — he will remain CEO and Des Traynor will continue running R&D — and praised the resources Salesforce would bring to accelerate their roadmap.

ServiceNow breaks CEO pledge with hundreds of layoffs across California

ServiceNow is cutting hundreds of employees, reversing years of CEO Bill McDermott's "no job cuts" promise as the company shifts its workforce strategy toward AI-driven efficiencies.

ServiceNow confirmed it is laying off a three-figure number of employees, with the most recent WARN notices covering 63 workers at its San Diego office and dozens more across California. The cuts disproportionately affect middle-market sales, solution consulting, training, and customer support roles, with roughly half of the eliminated positions being senior-level directors and senior managers. Internal estimates of the total layoff count range from 300 to 2,500.

The restructuring marks a significant turning point for the company, which employed approximately 29,000 people as of December 2025 and had previously vowed to absorb natural attrition through AI productivity gains. McDermott said during Q1 earnings that ServiceNow would "manage headcount with discipline to end the year where we started," framing the cuts as a reallocation toward AI-focused roles.

ServiceNow patches critical API flaw after unauthenticated access exposed customer data

An unauthenticated vulnerability in a ServiceNow API endpoint allowed unauthorized parties to query customer instance data, prompting an emergency security update on June 5, 2026.

ServiceNow applied the fix to hosted instances through advisory KB3067321, addressing a flaw in the `/api/now/related_list_edit/create` endpoint that shipped with `requires_authentication=false`. The vulnerability affected customers on the Australia platform release and others who had made certain configuration changes on earlier releases. ServiceNow detected anomalous activity and confirmed that a subset of customer instances were successfully queried, potentially exposing IT support tickets, employee records, and workflow data.

The company stated that the observed malicious activity was believed to be tied to security researchers or bug-bounty submissions rather than malicious threat actors, though it continued investigating at the time of disclosure. The flaw was first reported via ServiceNow's bug bounty program on April 22, 2026, and was not patched until June 5, creating a window of approximately two months.

ServiceNow reshapes ITOM licensing into two tiers with AI-driven operations focus

ServiceNow's April 2026 packaging release retired its legacy three-tier ITOM model in favor of ITOM Advanced and ITOM Prime, eliminating a standalone foundation tier and consolidating observability under the new structure.

ITOM Advanced focuses on CMDB health, agentless discovery, service mapping, and data integrity. ITOM Prime adds event management, log analytics, anomaly detection, synthetic monitoring, and twice the Workflow Data Fabric credits per Subscription Unit compared to Advanced. The former separate GenAI add-on SKU for ITOM is deprecated, with Now Assist for ITOM rolled directly into both tiers.

ServiceNow also introduced combined Service Ops Advanced and Service Ops SKUs, which merge ITSM fulfiller seats and ITOM Subscription Units onto a single credit meter with a unified true-up cycle, aimed at organizations whose IT operations and SRE teams share budgets and goals.

ServiceNow reports $3.77 billion in Q1 revenue but stock sinks 12% on growth concerns

ServiceNow posted strong Q1 2026 results with $3.77 billion in revenue, up 22 percent year over year, and $469 million in profit, but investors reacted negatively to guidance on AI monetization pace and margin compression.

The stock fell 12 percent after hours on April 22 as analysts pressed management on when ServiceNow's Now Assist AI product would deliver meaningful acceleration. CFO Gina Mastantuono noted Now Assist is projected to generate $1.5 billion by year-end, though this falls well short of the $5 billion-per-quarter AI revenue gains reported by rival AI labs. The company also guided operating margins down from 82.1 percent to 81.5 percent, adding pressure after the stock had already declined 14 percent on the earnings call itself. ServiceNow closed at $102.15 on the Friday following the earnings report, down sharply from its 52-week high of $211.48.