AI News for September 17, 2026
Browse any day on a calendar, install it to your home screen, read offline.Read in the app →-
1TechCrunch05:07
Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?
Anthropic and OpenAI plan to embed outside safety evaluators inside their labs, giving researchers a level of pre-release access they have not had before. Researchers welcome the access but argue it does not amount to oversight without transparency, structural independence, and eventually regulation. TechCrunch ran a companion piece arguing labs should first tighten the access their own agents have.
Read original ↗ -
2The Verge00:30
Claude comes for Gemini with its own take on Docs and Slides
Anthropic added Docs and Slides to Claude, letting users generate documents and presentations inside a chat and then export, edit, and share them. The company is also merging regular chats and Cowork into a single Claude interface, starting with Pro and Max subscribers. TechCrunch also covered the release.
Read original ↗ -
3The Verge01:00
Google will now let any AI agent run your smart home
Google is opening Google Home to third-party AI agents through a new MCP server, now in early access. Tools such as Claude and ChatGPT can control connected devices, review camera summaries, and query home activity in natural language. Adopting the Model Context Protocol at this scale gives agent developers a standard route into a large consumer device install base. TechCrunch also reported the launch.
Read original ↗ -
4The Verge01:20
Apple might make servers again to cash in on the AI rush
Citing The Information, The Verge reports Apple is preparing to re-enter the server business and may work with Nvidia on it, its first enterprise server since the Xserve line was retired in 2011. Ars Technica reports the machine would be packed with M-series Ultra chips and could arrive in 2029. If it ships, Apple would be selling AI compute rather than only consuming it.
Read original ↗ -
5TechCrunch21:00
Former Infosys chief's AI startup nabs another $53M
An AI startup founded by a former Infosys chief has added $53M to its seed round only weeks after the initial raise. The Palo Alto company says it has already signed multiple seven-figure enterprise contracts within months of launch. The quick follow-on reflects continued investor appetite for enterprise AI teams that arrive with existing distribution.
Read original ↗ -
6The Verge20:00
A brief history of AI executives calling for regulation
The Verge revisits the long record of AI executives publicly asking to be regulated, after a fresh round of such statements from OpenAI's Sam Altman, Anthropic's Dario Amodei, Google DeepMind's Demis Hassabis, Microsoft's Satya Nadella, and others. The piece notes how little of the earlier rhetoric translated into binding rules, which is the context for reading the current calls.
Read original ↗ -
7TechCrunch07:43
Al Gore says the real AI risk isn't data centers — it's what industry leaders are warning about
In an interview with TechCrunch, Al Gore said he is not especially worried about emissions from AI data centers, arguing the grid and efficiency trends are manageable. He is more concerned by the AI industry's own warnings about where the technology is heading. The comments cut against the framing of the current data center backlash.
Read original ↗ -
8TechCrunch18:34
Amazon launches Alexa+ in India with Hindi support
Amazon has launched its Alexa+ assistant in India with Hindi support, opening early access to all customers in the market. India is one of Alexa's largest user bases, making it a significant test of the rebuilt assistant outside English.
Read original ↗ -
9The Verge07:40
Snap is launching a new Specs AI tool, and it’s coming to iOS and Mac
Snap introduced Specs Intelligence, an assistant that connects to a user's other digital accounts to handle work tasks and track travel details, and it will also run on iOS and Mac. Snap is positioning it as an anticipatory service rather than a chatbot, putting it against Meta's Muse and Gemini's Spark. TechCrunch separately covered Snap's renewed pitch for its $2,200 Specs glasses.
Read original ↗ -
10The Verge04:40
The AI data center e-waste problem is huge — and getting bigger
A new report says e-waste from the AI buildout has been badly underestimated, projecting enough discarded hardware by 2050 to fill 23 million shipping containers. The estimate is far above earlier studies, largely because accelerator refresh cycles are short. The findings feed into policy debates over disposal rules for data center equipment.
Read original ↗ -
11The Verge22:45
The sexy AI-powered dating app scams are here
A security researcher analyzing a fraudulent dating app found it serving AI-generated personas that place live voice calls to users, complete with fabricated profiles. The setup shows how cheap real-time voice and image generation has made romance fraud easier to run at scale. It is an early look at consumer-facing abuse of the same models sold as assistants.
Read original ↗ -
12The Verge04:59
The 2.5-hour AI-generated Odyssey movie is 2.5 hours too long
The Verge reviews Odysseus: The Fall, a two-and-a-half-hour retelling of The Odyssey made entirely with AI, and finds it barely watchable. The film arrives after Christopher Nolan's version drew large audiences to the same source material. For anyone tracking generative video, it is a useful data point on how far feature-length coherence still has to go.
Read original ↗ -
13TechCrunch23:00
Robots are waiting for a ChatGPT moment: Nvidia’s Les Karpas explains why at TechCrunch Disrupt 2026
TechCrunch previewed a Disrupt 2026 session in which Nvidia's Les Karpas will discuss why robotics has yet to have its ChatGPT moment despite rapid progress in foundation models. The talk targets the gap between lab demos and everyday deployment.
Read original ↗ -
14TechCrunch11:30
Your startup’s next teammate might be an AI agent: Gusto, Insight Partners, and Leland explain what that changes at TechCrunch Disrupt 2026
TechCrunch previewed a Disrupt 2026 panel with Gusto, Insight Partners, and Leland on how early-stage companies are building teams where humans and AI agents work side by side. The session focuses on keeping speed, accountability, and culture intact as agents take on real work.
Read original ↗
