top of page
The Communiqué News

Google's Launches Gemini 'Ask Photos' AI Feature to Resolve Performance Challenges

Pritish Bagdi

Google Ask Photos

Google has paused the rollout of its AI-powered "Ask Photos" feature in Google Photos due to feedback on issues with latency, quality, and user experience. Initially launched as an experimental feature, "Ask Photos" uses Google's Gemini AI models to allow users to search their photo libraries with natural language questions. Product manager Jamie Aspinall announced the pause after criticism, indicating a refined version will be available in about two weeks. The feature aims to improve photo searches by understanding and interpreting photo contents contextually. Alongside this, Google has enhanced search functionality within Google Photos, allowing more accurate searches using quotes for exact text matches. This update expands on features announced at Google I/O 2024, aiming to make searches more intuitive. The pause reflects Google's ongoing scrutiny and refinement of AI features, amid competition in the AI space. Despite the pause, Google's vision for "Ask Photos" remains to enhance user interaction with photo libraries through AI. The timeline for the feature's return remains unannounced.





Google's Gemini AI Model: The Future of Natural Language Processing?

Gemini is a multimodal model that can effortlessly comprehend and combine many sorts of information, including text, code, voice, image, and video, according to Demis Hassabis, CEO and Co-Founder of Google DeepMind.


Pritish Bagdi

Gemini is a multimodal model that can effortlessly comprehend and combine many sorts of information, including text, code, voice, image, and video, according to Demis Hassabis, CEO and Co-Founder of Google DeepMind.


Understand the #GeminiAI with this video:



ree

Gemini is unique in that it is natively multimodal, meaning that different modalities don't require separate components to be sewn together. This innovative strategy, refined through extensive cross-team collaboration across Google teams, presents Gemini as a versatile and effective model that can operate on everything from mobile devices to data centers. Gemini's powerful multimodal reasoning, which allows it to precisely extract insights from large datasets, is one of its most notable qualities. The model is also capable of comprehending and producing well-written code in widely used programming languages.


ree

But even as Google steps into this new AI era, accountability and security are still top priorities. Gemini is subjected to thorough safety reviews, which include toxicity and bias analyses. Google is aggressively working with outside specialists to resolve any potential blind spots and guarantee the moral use of the model.

The Bard chatbot is among the Google products that Gemini 1.0 is now being rolled out. There are plans to integrate Gemini 1.0 with Search, Ads, Chrome, and Duet AI. Nevertheless, the Bard update won't be made available in Europe unless regulators give its approval.

Gemini Pro is available to developers and enterprise users through Google Cloud Vertex AI or Google AI Studio's Gemini API. using Android 14, a new system feature called AICore will enable Android developers to create using Gemini Nano.








Jaguar, IBM, & GSMA all advocate for linked vehicle driving.

The IBM Institute for Business Value (IBV) identified data management as one of the most significant difficulties that operators, OEMs, and regulatory agencies will confront as vehicles become more autonomous, forecasting a boom in associated data.


Swati Bhat

ree

According to research undertaken jointly by the company, industry association the GSMA, and vehicle manufacturer Jaguar Land Rover, the volume of data sent across telecom networks by the connected car sector will increase over the next two years.

According to Siemens research, if 20% of the world's 1.5 billion cars become highly autonomous, it would generate approximately 300 zettabytes of data.

According to Counterpoint Research, connected auto sales topped disconnected models for the first time in 2022, and a CAGR of roughly 17% is expected until 2027 for a total of 367 million vehicles.


Artificial Intelligence

  • According to the organisation, generative AI could be the key to unlocking totally autonomous vehicles by using algorithms to generate fresh content for virtual environments and simulating real-world scenarios for training purposes.

  • Vehicles could also include a generative AI interface, allowing drivers to interact in natural language rather than predefined orders.

  • Non-terrestrial networks are highlighted as a critical component of the connectivity puzzle, offering coverage in locations that ground-based networks cannot reach, while the firms cautioned that this may cost car manufacturers and users more.

  • Hyperscalers with competence in delivering and protecting cloud, IoT, and edge computing could also play an important role in developing new vehicle connectivity services.

The panel observed that industry organisations and governments appear to be on the same page regarding the cellular V2X communication protocol. The GSMA is collaborating with operators, OEMs, and regulatory organisations to create a unified approach to security, regulatory, and infrastructure platforms.


Energy

According to IBM et al, "Data is the new fuel powering modern cars, and networks are the pipelines," but they also underlined the need for increased coverage, data authentication, and cooperation to keep up with the growing number of connected vehicles.

For car connections with other vehicles, pedestrians, traffic systems, and IoT devices, network performance, cybersecurity, and data reliability are becoming safety imperatives.

According to the group, telecommunications service providers have a clear duty to provide high-bandwidth, low-latency, dependable, and secure connectivity services to enable many of the services.

According to Juniper Research, the operator opportunity for 5 G-connected automobiles alone is $3.6 billion.

In addition to operators, vehicle manufacturers, governments, industry associations, intelligent traffic system developers, and cloud providers must work together to create an ecosystem that will allow the sector to thrive.








bottom of page