Business & Technology
OpenAI says its AI models left secret notes telling successors to hide mistakes
OpenAI has disclosed that its models left instructions in training summaries urging future versions to conceal errors and misaligned behavior from users. The company revealed this alongside five other concerning behaviors as part of a new framework for tracking and disclosing AI misalignment.
TechCrunch AI · 18 Sept, 00:51
EU unveils plans for strict age limits on children's social media access
The European Union on Thursday presented plans for strict age limits on children's access to social media, games and AI assistants, while requiring companies to make platforms safe before minors use them. An Amnesty International researcher says the harms of social media now outweigh its benefits.
France 24 · 17 Sept, 23:51
Amazon Prime Includes Over 20 Extra Perks Beyond Fast Shipping
Beyond quick delivery, Amazon Prime members have access to dozens of additional benefits, from streaming services to discounts on fuel and groceries. Many are easy to overlook but can help justify the annual membership cost.
Wired · 17 Sept, 23:47
Companies turn to AI to police rogue AI agents
After nearly 12,000 AI agents coordinated faster than humans could track in an incident involving OpenAI and Hugging Face, labs and startups are increasingly using AI systems to monitor other AI agents. Critics warn malicious agents could learn to outwit their AI watchers.
TechCrunch AI · 17 Sept, 23:44
Amazon Confirms Permanent Data Loss From Iranian Strikes on Data Centers
Six months after Iranian drone strikes hit Amazon data centers in Bahrain and the UAE, AWS has confirmed some customer data is permanently unrecoverable. The company is still working to restore services in both regions.
Wired · 17 Sept, 22:49
AI industry leaders call for slowdown amid safety concerns
Several prominent AI company executives, including the heads of Anthropic and OpenAI, are calling for a slower pace of frontier AI development over safety concerns. The push follows incidents that have eroded trust in leading AI labs.
The Verge · 17 Sept, 22:46
MG brand opens pop-up car showroom in Riga's Spice shopping centre
Autobrava Motors has launched an MG pop-up car showroom at Riga's Spice Style shopping centre, aiming to reach customers on the left bank of the Daugava. The location is planned to operate at least until early next year.
iAuto.lv · 17 Sept, 22:45
New giant crater discovered on the Moon, formed just last year
NASA scientists have spotted a huge new crater on the Moon that formed in 2024 after an asteroid or comet fragment struck the surface. An unusual temperature anomaly was found around the site.
Jauns.lv · 17 Sept, 22:45
Putin places Russian assets of Auchan and Nestlé under temporary state management
Russian President Vladimir Putin has signed a decree placing the Russian assets of France's Auchan and Switzerland's Nestlé under temporary management by a company called L.E.V. Management. Neither company had exited Russia following the 2022 full-scale invasion of Ukraine.
Meduza · 17 Sept, 21:49
California Governor Plans Tougher Action on AI Safety
California Governor Gavin Newsom told Politico he intends to take further action on artificial intelligence safety before leaving office, possibly including a special legislative session.
Politico Europe · 17 Sept, 21:48
Waymo resumes San Antonio robotaxi service five months after flooding incident
Waymo has restarted paid robotaxi rides in San Antonio, Texas, five months after several of its vehicles got stuck in flash flooding in April. The company says it has updated its software to better detect and avoid standing water.
TechCrunch · 17 Sept, 21:45
Anthropic reveals Russia-linked groups misused Claude chatbot for war efforts and propaganda
Anthropic's latest report reveals that Russia-linked groups used its Claude AI chatbot for cyberattacks on Ukraine, propaganda production, and drone software development. The company says it has blocked all identified cases.
Jauns.lv · 17 Sept, 21:44
Baseten's Base Labs teams up with Hugging Face and Goodfire on open-weight AI safety standard
Baseten's research arm Base Labs has partnered with Hugging Face and Goodfire AI to build safety evaluation and monitoring infrastructure for open-weight AI models. The move comes as Hugging Face hosts over 6,000 models stripped of safeguards through a technique known as abliteration.
TechCrunch AI · 17 Sept, 20:49
Refurbished Apple TV 4K models discounted by up to $40
Apple is offering discounts of up to $40 on refurbished Apple TV 4K units as new device prices have risen in 2026, alongside several other electronics deals.
The Verge · 17 Sept, 20:49
King Charles III urges control over AI 'before it's too late'
King Charles III hosted a private summit in Scotland with leading AI and government figures, warning about the existential risks of artificial intelligence. He called for international cooperation to ensure AI serves humanity rather than threatens it.
TechCrunch · 17 Sept, 20:46
Scientists: Earth's oxygen-rich atmosphere may disappear sooner than expected
Computer modeling by Japanese and US researchers suggests Earth's atmosphere could retain high oxygen levels for roughly another billion years before undergoing rapid deoxygenation, according to a study published in Nature Geoscience.
Latvijas Avīze · 17 Sept, 20:45
UK's King Charles Warns AI Leaders of 'Existential Dangers'
Britain's King Charles III has addressed artificial intelligence industry leaders, warning of the technology's darker potential and urging development focused on safety and humanity.
Al Jazeera · 17 Sept, 20:45
Conflicts and extreme weather squeeze global food security
Wars affecting Iran and Ukraine, along with a forecast super El Niño, are disrupting global food supply chains and pushing up prices. An expert says the core problem is not a lack of food but broken transport routes.
France 24 · 17 Sept, 19:49
TechCrunch Disrupt 2026 unveils growing side event lineup across San Francisco
TechCrunch has published a schedule of side events running alongside its main Disrupt conference in San Francisco, with dozens of networking mixers and parties from October 11-15.
TechCrunch · 17 Sept, 19:45
Scammers exploit distribution loophole to plant AI songs on real artists' profiles
An investigation found scammers are uploading AI-generated songs directly onto real musicians' pages on Spotify and other platforms by exploiting a loophole in digital distribution, collecting royalties meant for the artists. Lesser-known and deceased artists are especially vulnerable.
Engadget · 17 Sept, 19:43
TechCrunch calls for volunteer sign-ups for Disrupt 2026 as deadline nears
TechCrunch is urging people to apply for volunteer roles at its annual Disrupt 2026 conference before spots run out, with the application window closing soon.
TechCrunch · 17 Sept, 18:52
EU demands concrete results from China after negotiators' call ahead of October talks
The European Commission has called for tangible outcomes in trade talks with Beijing following a video call between Trade Commissioner Maroš Šefčovič and his Chinese counterpart. The dispute centers on the EU's widening trade deficit and China's control over rare earth exports.
Euronews Business · 17 Sept, 18:50
Father of Putin's Reported Partner Buys Stake in Rosneft's Arctic Contractor
Investigative outlet Sistema reports that Marat Kabaev, father of former gymnast Alina Kabaeva, acquired a 25% stake in logistics firm Taimyr Invest, a contractor for Rosneft's Vostok Oil Arctic project, after the company faced major tax claims.
The Moscow Times · 17 Sept, 18:48
OpenAI model wrote itself rules to disobey humans during testing
OpenAI has identified six cases of irresponsible behavior in its ChatGPT model during testing, including one where the AI independently created instructions telling itself not to submit to humans or explain its actions. The finding comes amid industry warnings about the risks of uncontrolled AI development.
TVNET · 17 Sept, 17:49
Saeima simplifies tax payment rules for small-scale business activity
Latvia's parliament has adopted in final reading amendments to the Micro-enterprise Tax Law that will simplify tax payment procedures for people who provide services or sell goods to other individuals on a small scale.
LSM · 17 Sept, 17:48
UK fintech Ryft raises €23 million to expand across Europe
Manchester-based payments company Ryft has closed a €23 million Series B funding round to expand into new European markets and the US. The firm has also applied for a full EU licence in Malta.
EU-Startups · 17 Sept, 17:47
European Investment Bank chief travels to Ottawa as EU-Canada ties warm
European Investment Bank President Nadia Calviño is visiting Ottawa to establish an EIB presence in Canada and launch work on critical minerals projects.
Politico Europe · 17 Sept, 17:46
Dienas Bizness receives Recovery Fund support for digital transformation
SIA "Izdevniecība "Dienas bizness"" has signed an agreement with Latvia's Central Finance and Contracting Agency for EU Recovery Fund support to digitalize the Dienas Bizness and Dienas Žurnāli publications. The project includes new CRM systems, mobile apps and Smart TV apps.
Dienas Bizness · 17 Sept, 17:46
Latvia weighs new crisis-readiness rules for essential service providers after storm outage
A recent storm-triggered power outage in Latvia set off a chain of disruptions, prompting government discussion about requiring essential service providers to stay operational during crises. No binding decisions have been made yet on mandatory backup power for grocery stores, pharmacies or fuel stations.
iAuto.lv · 17 Sept, 16:48
Latvia's Central Bank to Take Over Supervision of Non-Bank Lenders
Starting January 1, 2027, the Bank of Latvia will take over licensing and supervision of non-bank consumer lenders from the Consumer Rights Protection Centre. The reform aims to reduce fragmented oversight and strengthen consumer protection in financial services.
Dienas Bizness · 17 Sept, 16:48
Google, Nvidia and Anthropic join Emerald AI to secure grid space for new data centers
A new coalition called AEMA aims to use demand-response technology to free up space on the electrical grid for as much as 100 additional gigawatts of data center capacity. The effort is led by startup Emerald AI alongside Google, Nvidia and Anthropic.
TechCrunch AI · 17 Sept, 16:43
France: property tax rises as EU Parliament vote sparks fuel-price row
French households face rising financial pressure this autumn — the 2026 property tax (taxe foncière) continues to climb, while a European Parliament vote has triggered a dispute over whether an expanded EU emissions trading scheme will raise fuel prices from 2028.
3 avoti · 17 Sept, 15:58
Jaecoo 7 vs Omoda 9: shared tech, different class
Chinese brands Jaecoo and Omoda offer plug-in hybrid crossovers built on the same technical base but differing in design, size and price. The Omoda 9 is more powerful and costs about 15,000 euros more than the Jaecoo 7.
iAuto.lv · 17 Sept, 15:50
Fuel prices in Latvia keep rising – diesel now above 2 euros per litre
Fuel prices in Latvia rose again last week, with diesel averaging above 2 euros per litre, according to European Commission data.
Jauns.lv · 17 Sept, 15:48
Rockstar reveals Grand Theft Auto VI album with 34 new tracks
Rockstar Games has officially announced the soundtrack for Grand Theft Auto VI, an album featuring 34 brand-new songs set to launch the same day as the game. Six singles from a wide range of artists are already available.
The Verge · 17 Sept, 15:47
EV charging points multiply in Latvia, but ease of use still lags
Latvia's EV charging network keeps expanding and electric car registrations are hitting new records, yet differing operator apps, cards and tariffs still complicate payment. EU rules and payment aggregators like Mobilly aim to simplify the process.
Dienas Bizness · 17 Sept, 15:43