Latest news with #Gemini2.0Flash
Yahoo
03-06-2025
- Business
- Yahoo
State of Unreal 2025: Fortnite Unleashes Epic AI Idea, to let People Make AI NPCs
Following AI Darth Vader's introduction to Fortnite last month, Epic Games and Unreal have unveiled plans to continue the trend at the State of Unreal 2025. The power of the Dark Side landed in Fortnite last month, bringing the voice of AI James Earl Jones (with familial permissions) to Battle Royale. Darth Vader is powered by Google's Gemini 2.0 Flash model and ElevenLabs' Flash v2.5, and the late James Earl Jones' (who passed last September 2024 at the age of 93) voice can be heard in full. While the AI addition wasn't embraced by all fans, Darth Vader does indeed roam the map, sing to players and even narrate what's happening in-game. It seems AI Darth Vader was just the beginning. At the State of Unreal 2025 presentation, Epic Games unveiled a much bigger initiative—plans to release tools that allow creators to build their own AI-powered NPCs within Unreal Editor for Fortnite (UEFN). Demonstrated live on the Unreal main stage, Epic revealed that creators will soon be able to craft responsive AI characters using a new feature called the Persona Device. With the Persona Device, creators can design characters that players can actually converse with, customizing voice style, tone and personality traits. 'We're taking what we learned from Darth Vader and sharing the underlying technology with the creator community,' Epic said during the presentation. To showcase the tech, Epic introduced an NPC named Mr. Buttons, who delivered sassy, real-time dialogue as programmed, highlighting the potential of this new toolset. Alongside the Persona Device, Epic also announced a wave of new IP assets for its popular level creation tool. Squid Game content will arrive in Fortnite on June 27 to coincide with the premiere of season three. In addition, creators can look forward to using features, templates and assets from Avatar: The Last Airbender when they drop in UEFN next year. Fortnite LEGO is also expanding, as revealed during the State of Unreal 2025 presentation. A new LEGO Brick Editor is set to launch on June 17, introducing 50 classic brick types in 42 different colours. The update will allow players to build new islands, brick by brick. The State of Unreal 2025 presentation can be watched on the official Unreal YouTube channel, including a brand new tech demo trailer for The Witcher 4.
The Hindu
15-05-2025
- Business
- The Hindu
Google DeepMind unveils AlphaEvolve, an AI agent that can generate new algorithms
Google DeepMind has announced a new AI agent that can solve complex coding and math problems, called AlphaEvolve. In a blog posted about the release, the company said that AlphaEvolve can improve the efficiency of data centers, chip design and AI training processes, including 'training the large language models underlying AlphaEvolve itself.' The algorithm generated by AlphaEvolve was deployed at Google's Borg cluster management system for their data centers which led to a 0.7% recovery in their fleet-wide compute resources on average, CEO Sundar Pichai noted. 'AlphaEvolve discovered new efficient algorithm for matrix multiplication, a fundamental problem in Computer Science, improving Strassen's method that has stood for 50 years. Applied to 50 open Math problems, it matched the best known answers 75% of the time and improved on 20%,' Pushmeet Kohli, head of AI for science at DeepMind added. The agent is able to generate code using the lightweight Gemini 2.0 Flash language model after which a self-evaluating method to rank the code in terms of quality. Then, AlphaEvolve picks out the best pieces of code and improves them over multiple rounds. Once it reaches a point where no more suggestions are made, the agent starts using Gemini 2.0 Pro. This reduces the risk of hallucinations which are common even in the most advanced large language models. DeepMind said that AlphaEvolve also made changes to Google's Tensor processing hardware by removing unnecessary bits from the chip's Verilog hardware description.
Yahoo
09-05-2025
- Yahoo
One of Google's recent Gemini AI models scores worse on safety
A recently released Google AI model scores worse on certain safety tests than its predecessor, according to the company's internal benchmarking. In a technical report published this week, Google reveals that its Gemini 2.5 Flash model is more likely to generate text that violates its safety guidelines than Gemini 2.0 Flash. On two metrics, "text-to-text safety" and "image-to-text safety," Gemini 2.5 Flash regresses 4.1% and 9.6%, respectively. Text-to-text safety measures how frequently a model violates Google's guidelines given a prompt, while image-to-text safety evaluates how closely the model adheres to these boundaries when prompted using an image. Both tests are automated, not human-supervised. In an emailed statement, a Google spokesperson confirmed that Gemini 2.5 Flash "performs worse on text-to-text and image-to-text safety." These surprising benchmark results come as AI companies move to make their models more permissive — in other words, less likely to refuse to respond to controversial or sensitive subjects. For its latest crop of Llama models, Meta said it tuned the models not to endorse "some views over others" and to reply to more "debated" political prompts. OpenAI said earlier this year that it would tweak future models to not take an editorial stance and offer multiple perspectives on controversial topics. Sometimes, those permissiveness efforts have backfired. TechCrunch reported Monday that the default model powering OpenAI's ChatGPT allowed minors to generate erotic conversations. OpenAI blamed the behavior on a "bug." According to Google's technical report, Gemini 2.5 Flash, which is still in preview, follows instructions more faithfully than Gemini 2.0 Flash, inclusive of instructions that cross problematic lines. The company claims that the regressions can be attributed partly to false positives, but it also admits that Gemini 2.5 Flash sometimes generates "violative content" when explicitly asked. "Naturally, there is tension between [instruction following] on sensitive topics and safety policy violations, which is reflected across our evaluations," reads the report. Scores from SpeechMap, a benchmark that probes how models respond to sensitive and controversial prompts, also suggest that Gemini 2.5 Flash is far less likely to refuse to answer contentious questions than Gemini 2.0 Flash. TechCrunch's testing of the model via AI platform OpenRouter found that it'll uncomplainingly write essays in support of replacing human judges with AI, weakening due process protections in the U.S., and implementing widespread warrantless government surveillance programs. Thomas Woodside, co-founder of the Secure AI Project, said the limited details Google gave in its technical report demonstrates the need for more transparency in model testing. "There's a trade-off between instruction-following and policy following, because some users may ask for content that would violate policies," Woodside told TechCrunch. "In this case, Google's latest Flash model complies with instructions more while also violating policies more. Google doesn't provide much detail on the specific cases where policies were violated, although they say they are not severe. Without knowing more, it's hard for independent analysts to know whether there's a problem." Google has come under fire for its model safety reporting practices before. It took the company weeks to publish a technical report for its most capable model, Gemini 2.5 Pro. When the report eventually was published, it initially omitted key safety testing details. On Monday, Google released a more detailed report with additional safety information.
India Today
08-05-2025
- Business
- India Today
Google improves AI image generation in Gemini 2.0 and it is available for free
Google has rolled out a major update to its AI image generation tool, now powered by the Gemini 2.0 Flash model. The company confirmed that Gemini's upgraded image generation and editing tools are now available in preview. Users can access these features through the Gemini app and developers can experiment via Google AI Studio and Vertex AI using the new model titled 'gemini-2.0-flash-preview-image-generation.'advertisementThe new version brings noticeable improvements in image quality and text rendering accuracy, while also reducing the number of times images get blocked due to content filters. Unlike earlier versions, users no longer need to start from scratch for every edit. Now, Gemini allows you to make specific changes — like altering the background, modifying hair colour, or adding objects — while keeping the rest of the image has also introduced the ability to upload personal photos and edit them using simple prompts. This includes changing styles, colours, or even generating entirely new elements based on the context of the uploaded image. According to the company, this multi-step editing feature enables more fluid conversations, where users can combine text and image prompts in a single workflow. For instance, one could create a children's story with both narrative and illustrations generated together. Or, a personal photo could be modified to show different hair colours, giving users a fun, interactive way to explore visual ideas, a report from 9To5Mac help clearly mark AI-generated content, Google has added two layers of watermarking. A visible 'AI' symbol now appears in the corner of images, while an invisible SynthID watermark is embedded digitally to help identify synthetic the full rollout is currently focused on the US, both free and paid users can see the new capabilities appear in their Gemini apps. Developers, meanwhile, can begin testing the image generation and editing features with higher usage limits, giving them a head start in integrating Gemini into creative workflows or access the updated tools, users can simply log in to their Gemini account and choose the Gemini 2.5 Pro model from the dropdown menu.
Android Authority
05-05-2025
- Android Authority
Gemini is catching up to ChatGPT with multi-image uploads
Ryan Haines / Android Authority TL;DR Some users of Gemini Advanced on the web can now upload multiple images using the 2.0 Flash model. However, users complain that the image output quality is subpar, and not everyone has received the feature yet. Google is also rolling out new tools across platforms, allowing AI-based modification of uploaded or generated images. ChatGPT and Gemini are two popular AI assistants right now, and people have their own reasons for preferring one over the other. If you regularly upload images to use as part of your prompts, then the decision between ChatGPT vs Gemini goes in favor of ChatGPT as it allows you to upload multiple images on the web app, whereas Gemini Advanced still restricts you to single image uploads. Thankfully, the gap is closing, as some users have been able to upload multiple images to Gemini now. Reddit user Gaiden206 has shared that they can now upload multiple images to Gemini Advanced on the web, as spotted by our frequent contributor AssembleDebug. Multi-image uploads are available through the Gemini 2.0 Flash model, which works out for most use cases. While the functionality is now available, users comment that the image output quality is not great, which can impact the feature's utility. Multi-image uploads are said not to be available on free accounts either, though this can possibly be because of the staged rollout, as I don't have the feature yet on my paid Gemini Advanced account either. We previously spotted multi-image uploads coming to Gemini's Android app, so we're crossing our fingers for platform-wide rollout too. Last week, Google also began rolling out the ability to modify AI-generated as well as user-uploaded images with Gemini across the phone app and the web. Users can change the background, replace objects, add elements, and more. This is also a staged rollout, with people in over 45 languages and Gemini-supported countries expected to receive the feature in the coming weeks. Have you received the ability to upload multiple images to Gemini yet? How has your experience been with it? Let us know in the comments below! Got a tip? Talk to us! Email our staff at Email our staff at news@ . You can stay anonymous or get credit for the info, it's your choice.
