Google's Bard levels up with Gemini: The new AI powerhouse taking on ChatGPT

Google has launched Gemini, its much anticipated cutting-edge artificial intelligence model, into its chatbot, Bard. Positioned to outshine competitors, including ChatGPT, Google's Gemini boasts superior reasoning abilities, heralding a new era in AI innovation. With OpenAI, the creator of ChatGPT, navigating recent upheavals and Google seizing the opportunity to surge ahead in the AI race, the tech giant is ready to take the lead in generative AI.

Sundar Pichai, Google's CEO, emphasised the groundbreaking nature of Gemini, labelling it as one of the company's most significant science and engineering endeavours to date. As Gemini proves its mettle by outperforming human experts in various benchmarks, including problem-solving, mathematics, and ethics.

Google unveils Gemini: A giant leap forward in AI

Alphabet, Google's parent company, introduced Gemini, its most sophisticated AI model to date. This unveiling marks a significant stride in the rapidly evolving artificial intelligence (AI) landscape, positioning Google to compete head-on with industry counterparts OpenAI and Meta in the quest for AI supremacy.

Gemini is the inaugural AI model to emerge from Alphabet's consolidated AI research units, DeepMind and Google Brain, which have merged into a unified division named Google DeepMind, steered by DeepMind CEO, ‘Demis Hassabis’.

Distinguished by its "multimodal" architecture, Gemini sets itself apart by seamlessly processing diverse data types, including text, code, audio, images, and video simultaneously.

Gemini’s three sizes of brilliance

The AI model comes in three sizes—Ultra for intricate tasks, Pro for versatile applications, and Nano designed for on-device functions.

Alphabet CEO, ‘Sundar Pichai’, expressed the company's ambitious vision, stating, "This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company."

Gemini Pro's early access for developers

Gemini Pro is set to be accessible to developers through the Gemini API in Google AI Studio and Google Cloud Vertex AI starting December 13. On the mobile front, Gemini Nano will be accessible to Android developers through AICore, a new capability introduced in Android 14, initially available on Pixel 8 Pro devices.

Selected customers, developers, partners, and safety experts will gain early access to Gemini Ultra for experimentation and feedback, with broader availability anticipated in the coming months.

Gemini integration across Google ecosystem

Google plans to deploy Gemini across its suite of products, with Bard, one of Google's language models, incorporating a fine-tuned version of Gemini Pro for advanced reasoning, planning, and understanding.

Gemini Nano is poised to enhance features on Pixel 8 Pro smartphones, such as the 'Summarise' function in the Recorder app and Smart Reply in Gboard, starting with WhatsApp and expanding to other messaging apps in the near future.

Additionally, Gemini contributes to optimising Google's generative AI search experience, Search Generative Experience (SGE), resulting in a 40 percent reduction in latency in the United States.

The power and flexibility of Gemini

Gemini Ultra's performance sets a new standard, outperforming human experts in 30 of 32 widely used academic benchmarks. Gemini Pro, before its public launch, surpassed GPT-3.5 in six of eight benchmarks, showcasing its prowess in language understanding and problem-solving.

According to Demis Hassabis, Gemini's multimodal reasoning capabilities make it adept at extracting insights from vast amounts of written and visual information. It excels in understanding complex subjects like math and physics, answering questions, and even generating high-quality code in popular programming languages.

Hassabis envisions Gemini as a flexible model capable of efficient operation across data centres and mobile devices. Google aims to continually enhance Gemini's capabilities, with plans to advance planning and memory functions and increase the context window for processing even more information.

Addressing potential risks with advanced protections

Recognising the transformative potential of Gemini, Alphabet is implementing new safeguards to address potential risks. Research into areas like cyber-offense, persuasion, and autonomy has been conducted, applying adversarial testing techniques to identify critical safety issues in advance of Gemini's deployment.

External experts and partners are collaborating with Google to conduct extensive stress-testing, ensuring that Gemini aligns with safety policies and AI principles.

As Alphabet dives into the Gemini era, the company is poised to shape the future of AI, leveraging the capabilities of this groundbreaking model to accelerate scientific discovery, foster innovation, and usher in a new era of technological progress.