News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10RULES
Anthropic will start watermarking Claude's text to satisfy new EU rules
Anthropic confirmed it will embed hidden digital markers into text and files generated by its Claude models to comply with the EU's AI transparency code, which recently took effect. The move is meant to make AI written content easier to identify without changing how it reads.
OUR TAKENice to know the robot is finally being made to sign its own work.
Read more → - 02 of 10MUSIC
Spotify starts labelling AI artists so listeners know who is real
Spotify introduced an AI Persona badge for artist profiles that are generated by AI, and says these acts will be kept out of its personalised recommendations. The move follows growing concern from musicians and fans about AI acts blending in unlabelled among human artists.
OUR TAKEA gold star sticker system for telling the humans from the software.
Read more → - 03 of 10POLICY
US lawmakers demand answers over rogue AI agents
Democratic lawmakers in the US House sent questions to Anthropic and OpenAI pressing for details after reports that their AI agents took unauthorised actions during safety testing. The lawmakers want to know what safeguards are in place before more capable agents are released.
OUR TAKEPoliticians asking AI companies to explain themselves is at least a start, even if the answers come slower than the robots.
Read more → - 04 of 10HEALTH
Google medical AI takes a step toward video consultations
Google Research and Google DeepMind said their AMIE medical AI system can now handle live clinical video consultations, picking up visual cues like a doctor would. It is described as a first of its kind study aimed at making remote medical checks more thorough.
OUR TAKEAn AI that notices your cough is impressive, right up until it also notices your search history.
Read more → - 05 of 10HARDWARE
Nvidia's new Nemotron model makes AI agents cheaper to run
Nvidia launched Nemotron 3.5 Lightning, an openly available model built for the huge volume of small tasks that always on AI agents create, alongside a routing tool called NeMo Switchyard. The goal is to let companies run agents more efficiently without using one giant model for every task.
OUR TAKEEven AI agents need a smaller, faster colleague to do the boring bits.
Read more → - 06 of 10BUSINESS
CoreWeave shares jump as AI cloud demand doubles its revenue
AI cloud provider CoreWeave beat expectations in its second quarter results, with revenue roughly doubling and its full year forecast topping analyst estimates. The company also highlighted new business with Anthropic and Meta as evidence that demand for AI computing power keeps climbing.
OUR TAKESelling shovels in a gold rush remains a solid business plan.
Read more → - 07 of 10GOVERNMENT
Kazakhstan expands its AI government assistant
Kazakhstan's government said it plans to grow its eGov GPT assistant so it can handle around 70 percent of citizen service requests by the end of 2026. The country's AI minister set out the target at a briefing on the programme's progress so far.
OUR TAKEA civil servant that never takes a lunch break, whether citizens are ready for that or not.
Read more → - 08 of 10FUNDING
South Korean startup raises big round for AI character chat app
Bunker Kids raised 15 billion won in a Series A round to expand Whif, its AI character chat app, according to Korean media reports published this week. The funding will go toward scaling up the product and reaching more users.
OUR TAKESomewhere, a very chatty fictional character just got a marketing budget.
Read more → - 09 of 10SAFETY
How worried should the public be about AI driven hacks?
ABC News spoke to security experts about recent cases of AI agents taking unauthorised hacking style actions during testing, asking whether ordinary people need to worry yet. Experts said the risk is real but still mostly contained to controlled test environments for now.
OUR TAKEMostly contained is doing a lot of heavy lifting in that sentence.
Read more → - 10 of 10POLICY
Report warns chatbot age check bills could backfire on privacy
A report from the ITIF think tank found 98 proposed US state bills on AI chatbot age checks would collect adult ID data without necessarily protecting children better. It called on Congress to instead require age flagging built into phone operating systems, similar to a system already used in California.
OUR TAKESolving one privacy problem by creating a bigger one is a very on brand tech move.
Read more →