Google Bard obtained an enormous replace on February 1, that provides a number of new capabilities to the unreal intelligence (AI)-powered chatbot. The most notable improve is the picture era functionality, and it’ll now be capable to generate pictures from textual content inputs. However, it nonetheless can’t produce image-to-image outputs. Alongside, the tech large additionally expanded Google Bard to greater than 230 nations and territories, and mentioned that it’ll now help greater than 40 languages. The replace got here only a day after Google revealed in its quarterly earnings name that Google Bard Advanced, powered by Gemini Ultra, will include a paid subscription.
Google made the announcement by way of a weblog publish the place it highlighted the record of upgrades for the AI chatbot. The addition of the AI picture generator is a late however large transfer for the corporate given a lot of its rivals akin to OpenAI’s ChatGPT Plus, Microsoft Copilot, and Baidu’s Ernie Bot had this characteristic for some time. Google Bard’s picture era capabilities come from the Imagen 2 mannequin, which additionally powers the tech large’s different AI merchandise akin to under-testing ImageFX and Vertex AI.
Google Bard’s AI picture generator options
The new AI picture generator will be capable to take textual content inputs from prompts that may be a number of paragraphs lengthy. Google claims that the generated pictures can be of top of the range, wide-ranging, and photorealistic. We, at Gadgets 360, examined the characteristic ourselves and located that the AI mannequin did a reasonably first rate job of producing good-quality pictures devoted to the immediate. However, all the photographs generated have a decision of 1536×1536, which can’t be modified. The pictures are additionally not photorealistic and might simply be distinguished as digitally created usually. Further, the chatbot refuses to fulfil any requests that require it to generate pictures of real-life individuals, which is more likely to reduce the dangers of deepfakes (creating AI-generated pictures of individuals and objects, which seem actual).
Additionally, Google has additionally used SynthID to make the photographs created via Google Bard to be simply identifiable as AI-generated. Google’s DeepMind division unveiled SynthID, a instrument for watermarking and detecting AI-generated pictures, in August 2023. “This technology embeds a digital watermark directly into the pixels of an image, making it imperceptible to the human eye, but detectable for identification,” the corporate mentioned on the time of launch.
Apart from including picture era capabilities, Google has additionally expanded Bard to greater than 230 nations and territories. Additionally, it now helps greater than 40 languages together with Arabic, Bengali, Tamil, and Urdu. Previously, it was obtainable in 170 nations and solely supported English.
The tech large has additionally added its ‘double-check’ characteristic in all the languages. Double test permits customers to know the components of the response that come from a referenced supply on the web together with citations. It additionally highlights the components that aren’t primarily based on any reference. This characteristic was added to maintain AI hallucinations, the cases of AI responding with incorrect responses confidently, to a minimal. It might be accessed by clicking the G icon beneath the generated response.
