More Gemini Features Coming to Google TV
Google TV integrates Gemini AI for image/video generation, enhances Google Photos, and adds a dedicated YouTube Shorts feed to the homepage.
Read moreIBM's foundational models are trained not only in language, but also in various forms of training, including code, time series data, tabular data, geospatial data, and IT event data.
According to a recent study, IBM is emerging as one of the major competitors benefiting from generative AI. IBM's generative AI product, Watsonx, has made this tech giant a significant beneficiary of this emerging technology. Geeta Gurnani, IBM's Chief Technology Officer and Technical Sales Director for India and South Asia, told AIM, "By using Watsonx, businesses can extend and accelerate the impact of advanced AI by accessing a complete technology stack for training, fine-tuning, and deploying AI models."
Generative AI gained attention quickly after the launch of ChatGPT due to its ability to engage in conversations almost like a human. However, IBM has been developing Watson for over a decade, initially designed as a question-answering computer system. Therefore, it can be said that IBM has prepared itself to bring generative AI capabilities to its clients. Gurnani stated that while many generative AI tools primarily focus on consumer and social purposes, Watsonx is specifically tailored for enterprises.
"Generative AI opens the door for improvement, increased productivity, and innovation across industries. Business leaders are excited about the prospect of leveraging foundational models and machine learning with their proprietary data to accelerate generative AI workloads. Watsonx enables them to deploy models applicable to various enterprise use cases, ranging from improving IT operations to enhancing HR functions," said Gurnani.
Watsonx
Through Watsonx, IBM provides its clients with an AI development studio that grants access to meticulously curated and trained foundational models, as well as open-source models. It also provides access to data repositories for data collection, cleaning, and training, along with tools for data and AI governance. "We have been at the forefront of working on foundational models, striving to deliver state-of-the-art foundational models to enterprises. We focus not only on the ability to generate the next word or image in a sequence but also on building and applying foundational models for completely unexplored business domains, such as geospatial intelligence, code, and IT operations," said Gurnani.
Furthermore, Gurnani further explained through an example of IBM's collaboration with NASA that IBM announced the development of large-scale geospatial AI models for NASA's Earth science satellite data and large language models for its Earth science literature. "This is the first time foundational models are being applied to NASA's satellite data. This has the potential to aid in estimating climate risks related to agriculture, monitoring forests for carbon offset programs, and developing predictive models to address and adapt to climate change," she said.
IBM's clients will be able to leverage open-source models on the Hugging Face platform, as well as IBM's developed foundational models. "The collaboration with Hugging Face allows IBM's clients to benefit from open-source models trained on accessible datasets in a secure environment, with compliance and proper data governance. This expands the range of available models and architectures, enabling clients to harness the best AI capabilities to meet their specific business needs," said Gurnani.
Driving Watsonx with Foundational Models
Through Watsonx, IBM's clients also gain access to IBM's foundational models, which are trained on a large-scale enterprise dataset, undergo robust filtering and cleaning processes, and have auditable data lineage. "These models are trained not only in language but also cover various forms such as code, time series data, tabular data, geospatial data, and IT event data," she said.
However, not everyone will have immediate access to these models. IBM plans to selectively provide access to these models to some clients, initially in the form of a beta release, and then make them available to all clients soon. IBM's foundational models are divided into three main parts. First is the model called fm.code, which aims to automate the productivity of developers and achieve automation of many IT tasks through a natural language interface.
Secondly, fm.NLP is a set of large language models designed for specific domains or industry sectors, utilizing curated data to mitigate biases more easily and allowing for rapid customization using client data. Lastly, fm.geospatial is a model built on climate and remote sensing data, helping organizations understand and plan for geophysical processes that may impact their business, such as climate patterns, biodiversity, land use, and more.
Mitigating Risks of Generative AI
While most businesses are eager to leverage generative capabilities, it also brings a range of challenges, such as biases. "At IBM, we are actively addressing challenges like biases and security while ensuring the ethical and responsible use of AI within Watsonx," said Gurnani.
According to Gurnani, IBM is reducing the risk of biases by enhancing generation through retrieval, which involves retrieving relevant data from a knowledge corpus before generating answers. "Users can also fine-tune existing models with domain-specific data to perform specific tasks, which also helps mitigate the risk of biases," she added.
Furthermore, she added that for any widely used technology, users must trust its outputs. Therefore, IBM is launching an AI governance toolkit, expected to be fully available later this year, which will help implement governance to mitigate risks, time, and costs associated with manual processes and provide the necessary documentation for driving transparency and explainable results. "It will also have mechanisms to protect customer privacy, proactively detect model biases and drift, and help organizations meet their ethical standards," she concluded.
Google TV integrates Gemini AI for image/video generation, enhances Google Photos, and adds a dedicated YouTube Shorts feed to the homepage.
Read moreAlphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Read moreAnthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Read moreApple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Read moreGoogle launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
Read moreYouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Read moreGoogle grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Read moreMistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.
Read moreCustom subliminal audio maker

AI Social Media Manager

AI Search Visibility

Workout and Calorie Tracker
Stock research platform

AI interview assistant

Scientific figure generator

AI courses with real code