Millions of US federal government workers are about to get a new AI assistant on their devices for free in the form of Microsoft Copilot. The move is part of a deal between Microsoft and the US General Services Administration (GSA) that’s also expected to save taxpayers $3.1 billion in its first year.

The centrepiece of this huge new agreement is a full year of Microsoft 365 Copilot at no extra cost for government workers using the high-security G5 licence. This is a push to get the latest AI tools into the hands of public servants quickly and safely, aiming to improve how the government operates.

Microsoft pushes the US government into the AI era

This deal aims to place the US government at the forefront of AI adoption. It’s a direct response to the administration’s AI Action Plan, designed to bring the power of modern artificial intelligence to everything from managing citizen enquiries to analysing complex data.

“OneGov represents a paradigm shift in federal procurement that is leading to immense cost savings, achieved by leveraging the purchasing power of the entire federal government,” explained FAS Commissioner Josh Gruenbaum.

The free Copilot offer is specifically for users on the Microsoft 365 G5 plan, the premium tier for departments that handle sensitive information and require the tightest security protocols. But the benefits extend further, with the deal helping agencies to use AI for automating routine tasks, freeing up people to focus on the work that matters most.

The agreement also makes it cheaper and easier for different departments to modernise their technology. By offering big discounts on Azure cloud services and getting rid of data transfer fees, it tackles a major headache that has often slowed down collaboration between agencies.

Security is not an afterthought

Of course, giving AI access to government systems raises immediate security questions. The deal addresses this head-on, with Microsoft emphasising that its core cloud and AI services have already passed FedRAMP High security authorisation, a critical standard for handling sensitive government data.

While the full FedRAMP High certification for Copilot itself is expected soon, it has already been given a provisional green light by the Department of Defense. The package also includes advanced security tools like Microsoft Sentinel and Entra ID to support the government’s “zero trust” security goal.

GSA Deputy Administrator Stephen Ehikian strongly encouraged government agencies to take advantage of the new tools.

“GSA is proud to partner with technology companies, like Microsoft, to advance AI adoption across the federal government, a key priority of the Trump Administration,” said Ehikian. “We urge our federal partners to leverage these agreements, providing government workers with transformative AI tools that streamline operations, cut costs, and enhance results.”

Helping government agencies to use AI effectively

Microsoft is also putting money into making sure the technology is actually used effectively. The company has committed an extra $20 million for support and training, including workshops to help agencies get the most out of the new tools and find other areas to reduce waste.

All told, the package is estimated to deliver more than $6 billion in value over the next three years.

“With this new agreement with the US General Services Administration, including a no-cost Microsoft 365 Copilot offer, we will help federal agencies use AI and digital technologies to improve citizen services, strengthen security, and save taxpayers more than $3 billion in the first year alone,” commented Satya Nadella, Chairman and CEO of Microsoft.

For the millions of people working within the US government, this agreement with Microsoft means that an AI-powered assistant is set to change their daily work.

See also: Marketing AI boom faces crisis of consumer trust

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.

Swarm Network, a platform developing decentralised protocols for AI agents, recently announced the successful results of its first Swarm, a tool (perhaps “organism” is the better term) built to tackle disinformation. Called Rollup News, the swarm is not an app, a software platform, nor a centralised algorithm. It is a decentralised collection of AI agents that collaborate to solve a bigger problem. The problem is that platforms like X allow any type of viral claims, some by incredibly influential people. How can we know what is true?

Currently, we try to solve this problem through equally loud opposing voices who offer facts or expert opinions. But if those sources are from a political side you oppose, why should you trust them? After all, these are people with their own motivations, and two additional issues are created: facts presented by a single person can easily get caught up in the “fake news” accusations; and misinformation presented as “facts” can be used to attack the ground truth.

Unfortunately, this isn’t just a current trend that will eventually lose its popularity and fade out. The more technology and access to varied news sources we have, the harder it becomes to not treat these sources equally. Some might be a traditional outlet that is legally liable if they falsify claims. Others might be a popular podcaster with an audience of millions, and whose fear-mongering ties nicely in with the products in their merch store. If it stopped at this, we could probably tell the truth from fiction. But it isn’t that simple. Official news channels have a history of spinning the news in their own bias, or ignoring other stories that are important to the public. On the other side, there are genuinely powerful influencers who seem to be hell bent on finding the truth and reporting it, no matter what side of the political spectrum it hits.

The world has become both confusing and dangerous, and the old “sticks and stones” saying has been proven false. After all, we have seen global elections swayed by disinformation, major policy shifts driven by false claims, lives damaged and lost as the result of powerful people lying, but lying loudly enough and often enough to sway large groups of people into believing them; and convincing these same groups that any facts to the contrary are the actual “fake news.”

Fixing fact checking

Given how challenging the disinformation industry is, and how absolutely slippery the truth is today, how can anyone hope to battle it? We have seen that people of all sides, realising that all news is skewed to some extent, will believe those sources that support their pre-existing beliefs.

A third party source, backed by overwhelming evidence, is needed to arbitrate. The source should not have an opinion, its methods should be transparent, and everyone should be able to see the same thing. This is nearly impossible, but the Web3 industry has shown that these attributes are what makes it incredibly powerful. Smart contracts handle billions in value daily, managing agreements from complete strangers from anywhere on the globe. The information is validated and the decisions are transparent, then locked in via the blockchain. The model has moved trillions of dollars using these very powerful, and neutral, tools.

Combine this trust with the other element Web3 excels in: decentralisation. Now attach another fast-emerging technology, the AI agent, which is easily built and designed to perform one task very well. The system is the centre of Swarm Network’s model, and its first deployment is Rollup News. The growing population of AI agents, the swarm, is designed to work collectively to scour the corners of X, find claims from users, and collectively test their validity using sources found in the information space. The results of these assessments are posted on the blockchain once validated by a large enough group of independent agents. Selective human participation helps to ensure the context and other subtle areas are handled well. The human element is also decentralised, preventing any particular viewpoint from being able to assert itself, and misconduct equals expulsion if someone tries to present fiction as fact. Rollup News has been operating for several months, with astonishing results: 128,000+ users have been onboarded, with over 5,000 rollup requests daily in July 2025. Over 3 million tweets were processed during that time, which is impressive in its own right, but when you consider the designed scalability of Web3 and AI agents working together, this is the linchpin of the battle in a world of disinformation.

The start of something new?

Rollup News’ success and Swarm Network’s larger model teach us a few things about fixing today’s problems. It is a demonstration that Web3 and AI are components in providing scalable solutions, that small AI agents can effectively work together to solve giant challenges, even if there is no centralised system. That decentralised environment, anchored by Web3, is the key to generating transparency, trust, and allowing strangers anywhere in the world to work together. Finally, the tokenisation of such a system creates the necessary incentives to attract more participants, fuelling the growth of a system. As long as it creates value, people will pay for its use, and those who help to validate and secure the decentralised network earn rewards. The type of truly free market system can scale up or down with the global demand faster than any traditional company. Swarm Network’s founder, Yannick Myson, sums it up nicely: “Rollup News shows what’s possible when AI agents, human insight, and blockchain converge. This isn’t a prototype – it’s working, and it’s scaling.”

We need to pay close attention to these lessons, as they offer a great deal of insight. First, the “truth-tech” sector, which is focused on using technology to combat mis/dis-information, has a strong blueprint for combining blockchain and AI. Second, there are many other sectors that need this level of global scaling and independent management, with untold value just ready to be developed and launched.

Image source: Unsplash

A team at Tencent’s Hunyuan lab has created a new AI, ‘Hunyuan Video-Foley,’ that finally brings lifelike audio to generated video. It’s designed to listen to videos and generate a high-quality soundtrack that’s perfectly in sync with the action on screen.

Ever watched an AI-generated video and felt like something was missing? The visuals might be stunning, but they often have an eerie silence that breaks the spell. In the film industry, the sound that fills that silence – the rustle of leaves, the clap of thunder, the clink of a glass – is called Foley art, and it’s a painstaking craft performed by experts.

Matching that level of detail is a huge challenge for AI. For years, automated systems have struggled to create believable sounds for videos.

How is Tencent solving the AI-generated audio for video problem?

One of the biggest reasons video-to-audio (V2A) models often fell short in the sound department was what the researchers call “modality imbalance”. Essentially, the AI was listening more to the text prompts it was given than it was watching the actual video.

For instance, if you gave a model a video of a busy beach with people walking and seagulls flying, but the text prompt only said “the sound of ocean waves,” you’d likely just get the sound of waves. The AI would completely ignore the footsteps in the sand and the calls of the birds, making the scene feel lifeless.

On top of that, the quality of the audio was often subpar, and there simply wasn’t enough high-quality video with sound to train the models effectively.

Tencent’s Hunyuan team tackled these problems from three different angles:

1. Tencent realised the AI needed a better education, so they built a massive, 100,000-hour library of video, audio, and text descriptions for it to learn from. They created an automated pipeline that filtered out low-quality content from the internet, getting rid of clips with long silences or compressed, fuzzy audio, ensuring the AI learned from the best possible material.

2. They designed a smarter architecture for the AI. Think of it like teaching the model to properly multitask. The system first pays incredibly close attention to the visual-audio link to get the timing just right—like matching the thump of a footstep to the exact moment a shoe hits the pavement. Once it has that timing locked down, it then incorporates the text prompt to understand the overall mood and context of the scene. This dual approach ensures the specific details of the video are never overlooked.

3. To guarantee the sound was high-quality, they used a training strategy called Representation Alignment (REPA). This is like having an expert audio engineer constantly looking over the AI’s shoulder during its training. It compares the AI’s work to features from a pre-trained, professional-grade audio model to guide it towards producing cleaner, richer, and more stable sound.

The results speak sound for themselves

When Tencent tested Hunyuan Video-Foley against other leading AI models, the audio results were clear. It wasn’t just that the computer-based metrics were better; human listeners consistently rated its output as higher quality, better matched to the video, and more accurately timed.

Across the board, the AI delivered improvements in making the sound match the on-screen action, both in terms of content and timing. The results across multiple evaluation datasets support this:

Tencent’s work helps to close the gap between silent AI videos and an immersive viewing experience with quality audio. It’s bringing the magic of Foley art to the world of automated content creation, which could be a powerful capability for filmmakers, animators, and creators everywhere.

See also: Google Vids gets AI avatars and image-to-video tools

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.