OpenAI is partnering with the US government to make its leading frontier models available to federal employees. Under the agreement, federal agencies can access OpenAI's models for 1 for the next year, per a Wednesday announcement from the company and the General Services Administration (GSA). The partnership is the culmination of months of effort on the part of OpenAI CEO Sam Altman and other OpenAI executives, who have been cozying up to the Trump administration since before President Donald Trump retook the White House in January. Since at least May of this year, high-ranking OpenAI employees have been meeting with the GSA and other government agencies, such as the Food and Drug Administration, to promote the company's tools, according to documents obtained by WIRED. On July 23, OpenAI chief operating officer Brad Lightcap and other OpenAI executives were invited to a private after-party hosted by the Hill and Valley Forum in Washington, DC.

OpenAI is reportedly in early talks about a sale of shares held by current and former employees that would value it at half a trillion dollars, overtaking Elon Musk's SpaceX. If the transaction goes ahead, the value of the ChatGPT developer would rise by about two-thirds, from 300bn ( 225bn). Musk's rocket companyis currently worth 350bn and is reportedly circling a 400bn price tag in a new fundraising. Bloomberg, which first reported the OpenAI talks, said existing investors, including Thrive Capital, have approached the company about buying employee shares. Other investors in OpenAI, which is based in San Francisco, include the Japanese investment company SoftBank, which led the 300bn financing, and Microsoft.

As the web fills up with junk text written by robots, it's no surprise that AI-generated slop is also making its way into ebooks, YouTube, and even Wikipedia. That last one is a particular problem, since Wikipedia's open access is its greatest strength and its primary method for dealing with misinformation. In response, the site's administrators are using new policies to combat the onslaught of AI text. According to a new policy write-up, Wikipedia administrators now have the authority and the tools to rapidly delete articles and edits that are obviously generated with large language models. But articles that are new or have been substantially rewritten can now be flagged with tags of obvious signs of LLM text, including phrases like "Here is your Wikipedia article on…" or citations and references to things that don't exist.

By the same token, Clune says, automating AI research and development could have enormous upsides. On our own, we humans might not be able to think up the innovations and improvements that will allow AI to one day tackle prodigious problems like cancer and climate change. For now, human ingenuity is still the primary engine of AI advancement; otherwise, Meta would hardly have made such exorbitant offers to attract researchers to its superintelligence lab. But AI is already contributing to its own development, and it's set to take even more of a role in the years to come. Here are five ways that AI is making itself better.

NVIDIA CEO and co-founder Jensen Huang commends President Donald Trump's A.I. agenda and outlines what the country's job future will look like on'Special Report.' FIRST ON FOX: The federal government is stepping into the future and embracing artificial intelligence, specifically ChatGPT, across its agencies, which proponents say will streamline productivity while solidifying President Donald Trump's pledge to keep the U.S. in the driver's seat of the cutting-edge technology, Fox News Digital exclusively learned. The U.S. General Services Administration announced Wednesday that OpenAI's ChatGPT Enterprise is now available to all federal agencies to incorporate into their workflow at a 1 per agency cost, the GSA told Fox Digital. The deal with OpenAI, the tech company behind ChatGPT, is part of GSA's OneGov Strategy that aims to modernize "how the federal government purchases goods and services" under the Trump administration. "The use of this tool has been deployed and tested with responsible policy makers, with responsible legal folks," GSA Federal Acquisition Service Commissioner Josh Gruenbaum told Fox News Digital of integrating AI into the federal government.

In a new apartment in Tel Aviv, the internet-connected lights go out. The smart shutters covering its four living room and kitchen windows start to roll up simultaneously. And a connected boiler is remotely turned on, ready to start warming up the stylish flat. The apartment's residents didn't trigger any of these actions. They didn't put their smart devices on a schedule.

The news: OpenAI has finally released its first open-weight large language models since 2019's GPT-2. Unlike the models available through OpenAI's web interface, these new open models can be freely downloaded, run, and even modified on laptops and other local devices. Why it matters: These releases re-establish OpenAI as a presence for users of open models. That's particularly notable at a time when Meta, which had previously dominated the American open-model landscape with its Llama models, may be reorienting toward closed releases--and when Chinese open models are becoming more popular than their American competitors. MIT Technology Review Narrated: AI means the end of internet search as we've known it The biggest change to the way search engines deliver information to us since the 1990s is happening right now.

The people who study nuclear war for a living are certain that artificial intelligence will soon power the deadly weapons. None of them are quite sure what, exactly, that means. In the middle of July, Nobel laureates gathered at the University of Chicago to listen to nuclear war experts talk about the end of the world. In closed sessions over two days, scientists, former government officials, and retired military personnel enlightened the laureates about the most devastating weapons ever created. The goal was to educate some of the most respected people in the world about one of the most horrifying weapons ever made and, at the end of it, have the laureates make policy recommendations to world leaders about how to avoid nuclear war.

Meta said WhatsApp "proactively detected and took down accounts before scam centres were able to operationalise them." In one case, WhatsApp worked with Meta and ChatGPT-developer OpenAI to disrupt scams linked to a Cambodian criminal group that offered cash for likes on social media posts to promote a fake rent-a-scooter pyramid scheme. It said scammers had used ChatGPT to create the instructions issued to potential victims. Typically, fraudsters would first contact potential targets with a text message before moving the conversation to social media or private messaging apps, said Meta. These scams were usually completed on payment or cryptocurrency platforms, it added.

The astonishing results of the transformer architecture on Natural Language Processing (NLP) tasks (Devlin et al. 2019; Radford et al. 2019), their scalation properties (Vaswani et al. 2017) and the massive amount of text data available (Wang et al. 2019; Foundation Accessed 27/05/2025) have led to the development of Large Language Models (LLM) whose performance towers above that of traditional Language Models (LM) (Zhang et al. 2021; BigScience et al. 2022). Furthermore, LLMs have been widely adopted for custom downstream tasks by leveraging the flexibility provided by fine-tuning (Chung et al. 2024) and their few-shot learning capabilities (Brown et al. 2020), establishing a new zeitgeist in the NLP community. These factors have led to their widespread adoption across major areas of society such as academia (Naveed et al. 2023; Meyer et al. 2023); industry, including sectors such as finance (Li et al. 2023), healthcare (Goyal et al. 2024) or law (Lai et al. 2024) and personal use, for example, as a personal assistant or search engine (Xiong et al. 2024; Microsoft Accessed 27/05/2025). Furthermore, the recent surge in their reasoning ability (Wei et al. 2022) and the development of cost-efficient models (Liu et al. 2024) suggest that there are still new avenues for improvement.