Gemini Live: What is it, and can it humanise generative AI?

Google has unveiled a new feature to let users talk to its generative artificial intelligence tool, Gemini, just as they would speak to a friend or an expert over the phone.

The new tool, Gemini Live, will allow users to have a back-and-forth conversation, where they can ask questions, brainstorm strategies, share ideas and even get advice in a natural and interactive way, the Alphabet-owned company announced on Tuesday.

The significance of the new feature lies in its ability to transform AI from a static or one-way conversational tool into a more integrated and responsive assistant.

Users can pause the conversation and return to it later, or interrupt to delve deeper into specific points of interest.

It is also available hands-free. Users can keep talking with the Gemini app in the background or even when their phone is locked, like a regular phone call.

“Gemini is becoming so much more than just a chatbot,” Amar Subramanya, vice president for engineering at Google Gemini, told media while introducing the feature.

“It is your well-trained personal assistant that is getting smarter day by day. It is fundamentally a different way of providing help to users in the moments that matter to them.”

Google said that while AI unlocks powerful new capabilities, it sometimes behaves in “unexpected ways or provides inaccurate information”.

Mr Subramanya said Google is working to minimise such cases by integrating user feedback and constantly evolving the technology.

The company, based in California, is competing with Microsoft-backed Bing and ChatGPT for a greater share of the generative AI market.

It had originally launched Gemini, previously known as Bard, in February last year.

In December, Google launched its first multimodal model, Gemini 1.0, in three sizes – ultra, pro and nano.

That was followed by an enhanced version, 1.5 Pro, with a one million-token context window. The latest 1.5 Flash is trained on up to a two million-token context window.

In natural language processing, a token refers to a single word or character.

Despite serious efforts and huge investment, Gemini has yet to show any significant contribution to the company's sales.

Google services business accounted for nearly 87.2 per cent of the company’s total sales in the second quarter of this year.

It added nearly $73.9 billion to overall revenue, nearly 11.5 per cent more on an annual basis. The company did not disclose the contribution of Gemini.

Who has access to it?

Gemini Live, first unveiled at the Google I/O conference in May, was introduced on Tuesday. At first, it will be available only in English and on Android smartphones.

The company said it plans to expand it to Apple’s iOS operating system and in more languages in coming weeks.

The company is trying to make interactions feel more natural and personalised.

It is offering Gemini Live in 10 different voices, with users able to select the one that best suits their preferences in terms of tone and style.

How much will it cost?

Gemini Live is available to users who have subscribed to the paid Gemini Advanced subscription.

To gain access to Gemini Advanced, users need to subscribe to the Google One AI premium plan that costs $19.99 a month.

It is available in 150 countries, including the US, the UK, the UAE and Saudi Arabia.

Users can sign in with their personal Google account. They must meet minimum age requirement of 18 years or above to use Gemini Advanced.

Growing competition in generative AI market

The introduction of Gemini Live marked an important moment in the competitive generative AI market, particularly in Google’s rivalry with OpenAI's ChatGPT.

In May, OpenAI launched GPT-4o, an updated version of the underlying large language model technology that powers ChatGPT.

Google aims to differentiate Gemini Live from the latest version of ChatGPT by emphasising natural interaction and personalisation.

Its ability to offer real-time, conversational assistance, combined with customisable voices, places it as a formidable competitor in the AI assistant market. For users, this means more customised experiences.

Regional entities are also taking aim at the leaders. In May, Abu Dhabi's Technology Innovation Institute introduced the second version of its large language model, Falcon 2, to compete with models developed by Google and OpenAI.

In the same month, Core42, a unit of Abu Dhabi's AI and cloud company G42, launched a bilingual Arabic and English chatbot developed in the UAE, Jais Chat.

And then, there's xAI's Grok, from OpenAI's friend-turned-enemy, Elon Musk.

Grok's current version, Grok-1.5, is available only to subscribers of X's Premium+ tier, which starts at $16 a month, or $168 a year.