In a groundbreaking update on February 1, Google Bard, the advanced AI-powered chatbot, marked a significant leap in its capabilities. The star feature of this update is the introduction of an AI image generation capability, reinforcing Google Bard’s position in the competitive landscape of conversational AI. However, the innovation doesn’t stop there; the tech giant also expanded Google Bard’s reach to over 230 countries, bolstering its language support to more than 40. Notably, this update closely followed Google’s announcement of the impending launch of Google Bard Advanced, fueled by the potent Gemini Ultra, featuring a paid subscription model.
AI Image Generation: A Late but Impactful Arrival
The highlight of Google Bard’s latest update is undeniably its newfound ability to generate images from textual inputs. Unlike its competitors such as OpenAI’s ChatGPT Plus, Microsoft Copilot, and Baidu’s Ernie Bot, Google Bard was relatively late to incorporate this feature. Powering this image generation capability is the Imagen 2 model, the same powerhouse behind Google’s other AI products like the in-development ImageFX and Vertex AI.
Features of Google Bard’s AI Image Generator:
- Text Input Versatility: The AI image generator can process prompts spanning multiple paragraphs, providing versatility in generating images.
- High-Quality Output: Google claims that the generated images exhibit high quality, covering a broad spectrum of visual content and aiming for photorealism.
- Limitations in Resolution: All images produced have a fixed resolution of 1536×1536, offering no flexibility in altering image dimensions.
- Distinctive Digital Creation: While the generated images are of good quality, they can often be identified as digitally created, lacking complete photorealism.
- Restrictions on Real-Life People: Google Bard refrains from generating images of real-life individuals, a proactive measure to mitigate the risks associated with deepfakes.
SynthID: Ensuring Authenticity in AI-Generated Images
To address concerns about the authenticity of AI-generated content, Google incorporated SynthID technology. SynthID, developed by Google’s DeepMind division, functions as a watermarking and detection tool specifically designed for AI-generated images. The digital watermark is imperceptible to the human eye but serves as an identification marker, establishing the authenticity of the images produced through Google Bard.
Global Expansion and Language Support
In tandem with enhancing its capabilities, Google Bard has expanded its global presence and language support. Previously accessible in 170 countries and limited to English, Google Bard now reaches users in over 230 countries and territories. Language support has been broadened to encompass more than 40 languages, including Arabic, Bengali, Tamil, and Urdu, reflecting Google’s commitment to inclusivity and accessibility.
Double-Check Feature: A Safeguard Against AI Hallucinations
To ensure the accuracy and reliability of responses, Google introduced the ‘double-check’ feature across all supported languages. This feature enables users to discern the origin of responses, indicating whether they are derived from an internet source with citations or purely generated by the AI. By mitigating AI hallucinations – instances where AI confidently provides inaccurate responses – Google Bard aims to uphold the integrity of its interactions.