The adoption of artificial intelligence (AI) generates both expectations and uncertainties. An IBM survey reveals that almost half of company managers plan to introduce it into their processes and that the same proportion fears its consequences and the difficulties in implementing it. The multinational announced today, during its annual meeting Think in the United States, the latest advances to banish these fears. Their applications have reached companies such as Meta, which incorporates these developments into the third version of Llama, Amazon, Microsoft, Mistral or SAP, among others, but they are far from the agents presented this week by Google and Open AI.
IBM wants to be a safety net in a dizzying scenario of artificial intelligence growth, which is advancing faster than the business capacity to understand and apply it without risks. To this end, the multinational has updated its comprehensive open source platform (Watsonx), which incorporates advances in the Granite and InstructLab models.
“We are opening the code of large Granite language models that IBM has created. And the reason this is incredibly important is because this is the best path to providing transparency, innovation and security to scale AI with an enterprise context. “We launched the highest-performing code models in their class in the world,” says the IBM vice president and director of the research division (IBM Research). Dario Gil.
“Open means choice. Open means more eyes on the code, more minds on the problems, and more hands on the solutions. For any technology to gain speed and become ubiquitous, three things must be balanced: competition, innovation and security. Open source is an excellent way to achieve all three,” completes the company’s CEO. Arvind Krishna.
With this strategy, the company has advanced this Tuesday new developments of Watsonx, the comprehensive artificial intelligence platform that IBM created a year ago to help companies scale and accelerate the implementation of AI solutions, from the creation and training of models from AI to its implementation and management in production.
This platform includes improvements in Granite (available today under Apache 2.0 licenses on HuggingFace and Github) to offer already trained models from the cloud and available to be incorporated into business applications without having to start from scratch and with security guarantees. “We launched them with an Apache 2.0 license to give the maximum level of flexibility to the community and our customers to build on them,” explains Gil.
IBM has also presented the InstructLab projectthe platform developed in collaboration with Red Hat to simplify the process of creating, training and deploying artificial intelligence (AI) and machine learning models (machine learning) through the use of generative AI.
“InstructLab allows our clients to be able to add knowledge, skills and their own data in a secure and reliable way for enterprise use cases,” explains Gil. IBM plans to leverage collaborative advances with open source to deliver additional value through integrations with watsonx.ai and the new Red Hat Enterprise Linux AI (RHEL AI) solution.
The AI models proposed by IBM use a lower range of parameters than systems from other companies. Gil explains it. “It is the industry trend: obtain the performance that is needed for the use cases and at an affordable cost.”
Along the same lines, the assistants built with this company’s systems do not try to reach the developments of the agents presented last week by Google and Open Ai. It is not your goal. Robert Thomas, head of sales and programming, argues that the developments of the other internet giants go to “very different markets” than the one IBM is looking for. “The focus I see at Open AI and at Google is interaction with the consumer. “This kind of thing is not our focus, but business use cases, digital work, which is automating repetitive tasks in an organization that then extends to things like data and AI governance,” he adds.
IBM’s strategy responds to forecasts that generative AI will power up to 1 billion applications by 2028 in order to automate processes. In this field, a tool (Concert) has been presented that will be generally available in June 2024 and that aspires to become the “nerve center” of the company’s information technologies. It will allow you to identify, predict and solve problems and risks, according to IBM.
Amazon Web Service (AWS) has been incorporated into the Watsonx ecosystem for the governance of artificial intelligence in its predictive models of machine learning and generative AI, Adobe and Meta, which has incorporated IBM technology and programming into Llama 3, the latest version of artificial intelligence with which Mark Zuckerberg’s company tries to follow Google and Open AI. For its part, Watsonx is now available to run on Microsoft Azure and Palo Alto Networks, the cybersecurity company that has risen to the top in billing. Salesforce, Mistral and SAP join the list of entities linked to the IBM ecosystem, as announced by the company.
You can follow EL PAÍS Technology in Facebook and x or sign up here to receive our weekly newsletter.
Subscribe to continue reading
Read without limits
_
#IBM #offers #roadblock #boost #development #artificial #intelligence