Latest news with #PushmeetKohli
Ammon
5 days ago
- Ammon
Google has a new tool to help detect AI-generated content
Ammon News - Google announced a new SynthID Detector tool at Google I/O that lets you check if content has been made with the assistance of Google's AI tools. In a blog post, Google DeepMind's Pushmeet Kohli describes SynthID Detector as 'a verification portal' that can 'quickly and efficiently identify AI-generated content made with Google AI.' It's also able to 'highlight which parts of the content are more likely to have been watermarked with SynthID.' SynthID watermarks are applied to AI-generated images, text, audio, and videos, including content generated by Google's Gemini, Imagen, Lyria, and Veo models, Kohli says. Here's how the tool works, according to Kohli: "When you upload an image, audio track, video or piece of text created using Google's AI tools, the portal will scan the media for a SynthID watermark. If a watermark is detected, the portal will highlight specific portions of the content most likely to be watermarked. For audio, the portal pinpoints specific segments where a SynthID watermark is detected, and for images, it indicates areas where a watermark is most likely." Google is starting to roll out the tool to 'early testers,' Kohli says in the post. 'Following the initial testing phase, the portal will gradually be rolled out to users who sign up to the waitlist to gain access to the SynthID Detector,' Kohli tells The Verge. 'We will take learnings from this cohort of professionals and work to implement content transparency more broadly.' I'm on the waitlist, but I haven't tested the tool myself, so I can't vouch for how well it might work. And will people actually use it when it's widely available? I hope so, but we'll have to wait and see. The Verge
Techday NZ
22-05-2025
- Techday NZ
Google unveils SynthID Detector to verify AI-generated content
Google has introduced the SynthID Detector portal, a tool designed to help users verify whether content has been generated using Google AI by detecting SynthID watermarks. The SynthID Detector provides a centralised platform where users can upload images, audio, video, or text to determine if the material contains a SynthID watermark. The portal is intended to address growing concerns around content authenticity, context, and verification as generative AI technologies continue to develop and become more widespread. Pushmeet Kohli, Vice President of Science and Strategic Initiatives at Google DeepMind, commented, "Advances in generative AI are making it possible for people to create content in entirely new ways — from text to high quality audio, images and videos. As these capabilities advance and become more broadly available, questions of authenticity, context and verification emerge." Kohli added, "Today we're announcing SynthID Detector, a verification portal to quickly and efficiently identify AI-generated content made with Google AI. The portal provides detection capabilities across different modalities in one place and provides essential transparency in the rapidly evolving landscape of generative media. It can also highlight which parts of the content are more likely to have been watermarked with SynthID." The SynthID watermarking technology was initially developed to embed imperceptible watermarks in AI-generated images to help mitigate misinformation and misattribution. These watermarks are designed to remain detectable even when the content is shared or altered. Google states that over 10 billion pieces of content have already been watermarked with SynthID since the system's launch. SynthID's coverage has since expanded beyond imagery to include text, audio, and video content generated by models such as Gemini, Imagen, Lyria, and Veo. Kohli explained, "When we launched SynthID — a state-of-the-art tool that embeds imperceptible watermarks and enables the identification of AI-generated content — our aim was to provide a suite of novel technical solutions to help minimise misinformation and misattribution. SynthID not only preserves the content's quality, it acts as a robust watermark that remains detectable even when the content is shared or undergoes a range of transformations. While originally focused on AI-generated imagery only, we've since expanded SynthID to cover AI-generated text, audio and video content, including content generated by our Gemini, Imagen, Lyria and Veo models across Google. Over 10 billion pieces of content have already been watermarked with SynthID." To use the SynthID Detector, users upload an image, audio track, video, or text generated using Google's tools to the portal. The system scans the content for SynthID watermarks, highlighting particular segments most likely to be watermarked. For audio, specific time segments are indicated, while images are displayed with areas that likely contain the watermark. The operational process is described as follows: "Upload content: Upload an image, audio track, video or piece of text created using Google's AI tools. Scan for watermarks: The portal scans the uploaded media and detects whether the uploaded content, or specific portions of it, contain a SynthID watermark. View results: The portal will present the results and if a SynthID watermark is detected, highlight which parts of the content are more likely to have been watermarked." The SynthID Detector is currently being rolled out to early testers before a wider release. Google notes that journalists, media professionals, and researchers are eligible to join a waitlist for early access to the portal. Google has also made the SynthID text watermarking framework open source. This enables developers to integrate the technology into their own models and projects. In addition to open sourcing, Google has begun to build partnerships with other companies to expand the reach and adoption of SynthID. Earlier this year, Google entered a partnership with NVIDIA to embed SynthID watermarks into videos generated by NVIDIA's CosmosTM preview NIM microservice. This collaboration increases the volume of content across the internet bearing the SynthID marker, broadening transparency beyond content produced solely by Google's AI models. Google has also announced a new partnership with GetReal Security, a company specialising in content verification, to further the capabilities of SynthID detection by external parties. The partnership is intended to support the detection of SynthID watermarks across a wider range of media and stakeholders. On the challenge of content transparency in the context of generative AI, Kohli stated, "Content transparency remains a complex challenge. To continue to inform and empower people engaging with AI-generated content, we believe it's vital to continue collaborating with the AI community and broaden access to transparency tools." Kohli also acknowledged the contributions of Miguel de Andrés-Clavera and the Google Partner Innovation team in developing the SynthID Detector portal.
The Verge
21-05-2025
- The Verge
Google has a new tool to help detect AI-generated content
Google announced a new SynthID Detector tool at Google I/O that lets you check if content has been made with the assistance of Google's AI tools. In a blog post, Google DeepMind's Pushmeet Kohli describes SynthID Detector as 'a verification portal' that can 'quickly and efficiently identify AI-generated content made with Google AI.' It's also able to 'highlight which parts of the content are more likely to have been watermarked with SynthID.' SynthID watermarks are applied to AI-generated images, text, audio, and videos, including content generated by Google's Gemini, Imagen, Lyria, and Veo models, Kohli says. Here's how the tool works, according to Kohli: When you upload an image, audio track, video or piece of text created using Google's AI tools, the portal will scan the media for a SynthID watermark. If a watermark is detected, the portal will highlight specific portions of the content most likely to be watermarked. For audio, the portal pinpoints specific segments where a SynthID watermark is detected, and for images, it indicates areas where a watermark is most likely. Google is starting to roll out the tool to 'early testers,' Kohli says in the post. 'Following the initial testing phase, the portal will gradually be rolled out to users who sign up to the waitlist to gain access to the SynthID Detector,' Kohli tells The Verge. 'We will take learnings from this cohort of professionals and work to implement content transparency more broadly.' I'm on the waitlist, but I haven't tested the tool myself, so I can't vouch for how well it might work. And will people actually use it when it's widely available? I hope so, but we'll have to wait and see.
Scientific American
15-05-2025
- Science
- Scientific American
New Google AI Chatbot Tackles Complex Math and Science
Google DeepMind has used chatbot models to come up with solutions to major problems in mathematics and computer science. The system, called AlphaEvolve, combines the creativity of a large language model (LLM) with algorithms that can scrutinize the model's suggestions to filter and improve solutions. It was described in a white paper released by the company on 14 May. 'The paper is quite spectacular,' says Mario Krenn, who leads the Artificial Scientist Lab at the Max Planck Institute for the Science of Light in Erlangen, Germany. 'I think AlphaEvolve is the first successful demonstration of new discoveries based on general-purpose LLMs.' On supporting science journalism If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today. As well as using the system to discover solutions to open maths problems, DeepMind has already applied the artificial intelligence (AI) technique to its own practical challenges, says Pushmeet Kohli, head of science at the firm in London. AlphaEvolve has helped to improve the design of the company's next generation of tensor processing units — computing chips developed specially for AI — and has found a way to more efficiently exploit Google's worldwide computing capacity, saving 0.7% of total resources. 'It has had substantial impact,' says Kohli. General-purpose AI Most of the successful applications of AI in science so far — including the protein-designing tool AlphaFold — have involved a learning algorithm that was hand-crafted for its task, says Krenn. But AlphaEvolve is general-purpose, tapping the abilities of LLMs to generate code to solve problems in a wide range of domains. DeepMind describes AlphaEvolve as an 'agent', because it involves using interacting AI models. But it targets a different point in the scientific process from many other 'agentic' AI science systems, which have been used to review the literature and suggest hypotheses. AlphaEvolve is based on the firm's Gemini family of LLMs. Each task starts with the user inputting a question, criteria for evaluation and a suggested solution, for which the LLM proposes hundreds or thousands of modifications. An 'evaluator' algorithm then assesses the modifications against the metrics for a good solution (for example, in the task of assigning Google's computing jobs, researchers want to waste fewer resources). On the basis of which solutions are judged to be the best, the LLM suggests fresh ideas and over time the system evolves a population of stronger algorithms, says Matej Balog, an AI scientist at DeepMind who co-led the research. 'We explore this diverse set of possibilities of how the problem can be solved,' he says. AlphaEvolve builds on the firm's FunSearch system, which in 2023 was shown to use a similar evolutionary approach to outdo humans in unsolved problems in maths. Compared with FunSearch, AlphaEvolve can handle much larger pieces of code and tackle more complex algorithms across a wide range of scientific domains, says Balog. DeepMind says that AlphaEvolve has come up with a way to perform a calculation, known as matrix multiplication, that in some cases is faster than the fastest-known method, which was developed by German mathematician Volker Strassen in 1969. Such calculations involve multiplying numbers in grids and are used to train neural networks. Despite being general-purpose, AlphaEvolve outperformed AlphaTensor, an AI tool described by the firm in 2022 and designed specifically for matrix mechanics. The approach could be used to tackle optimization problems, says Krenn, or anywhere in science where there are concrete metrics, or simulations, to evaluate what makes a good solution. This could include designing new microscopes, telescope or even materials, he adds. Narrow applications In mathematics, AlphaEvolve seems to allow significant speed-ups in tackling some problems, says Simon Frieder, a mathematician and AI researcher at the University of Oxford, UK. But it will probably be applied only to the 'narrow slice' of tasks that can be presented as problems to be solved through code, he says. Other researchers are reserving judgement about the tool's usefulness until has been trialled outside DeepMind. 'Until the systems have been tested by a broader community, I would stay sceptical and take the reported results with a grain of salt,' says Huan Sun, an AI researcher at the Ohio State University in Columbus. Frieder says he will wait until an open-source version is recreated by researchers, rather than a rely on DeepMind's proprietary system, which could be withdrawn or changed. Although AlphaEvolve requires less computing power to run than AlphaTensor, it is still too resource-intensive to be made freely available on DeepMind's servers, says Kohli. But the company hopes that announcing the system will encourage researchers to suggest areas of science in which to apply AlphaEvolve. 'We are definitely committed to make sure that the most people in the scientific community get access to it,' says Kohli.
