A RedMonk Conversation: IBM and embeddable AI - natural language processing and speech recognition.
Science & Technology
Introduction
In a recent conversation, James Governor, co-founder of RedMonk, spoke with Bill Higgins, a distinguished engineer and director of engineering at IBM, about the company’s advancements in foundational AI technologies, particularly focusing on natural language processing (NLP) and speech recognition. The discussion highlighted IBM’s strategy in making sophisticated AI technologies more accessible to software developers and companies.
The Journey of AI at IBM
Bill began by recounting the history of IBM’s AI endeavors, notably starting with the Watson Jeopardy system in 2011. This period marked a pivotal transition where AI evolved from being a distant concept to a practical tool that companies could adopt. Following the significant interest generated by Watson's success, IBM experienced a strong drive towards exploring various AI applications across industries.
In 2018, IBM acquired Red Hat and shifted towards a hybrid cloud and AI strategy. Bill played a crucial role in rolling out new development tools internally, which helped improve collaboration among developers. The goal was to embed AI into all of its products to enhance their capabilities and serve as references for foundational AI technologies.
As a result of that internal maturity, IBM made the decision to commercialize its AI technologies, recognizing that many software companies lack the extensive resources to develop cutting-edge AI in-house.
The Launch of Embeddable AI
Bill explained that IBM has developed around ten core AI components that have been rigorously tested and adopted internally. The latest announcements included the release of embeddable versions of IBM’s NLP and speech technology. These technologies, which have longstanding histories at IBM, are being made available to external software companies looking to enhance their applications with AI capabilities.
Target Audience
The primary audience for these embeddable AI technologies includes a wide range of software companies—not just traditional software service providers but also cloud companies that are seeking to implement enhanced capabilities in their products. As the software landscape evolves, IBM aims to cater to the needs of approximately 30,000 commercial software enterprises in North America alone.
Natural Language Processing: A Growing Trend
Natural language processing is currently one of the hottest areas in AI due to its vast applications and the practical feasibility of implementation. Bill noted that while NLP has been around for decades, it is now more accessible due to advancements in technology and the rise of cloud computing.
He pointed out that software leaders often need to recognize the value and feasibility of implementing AI in their products. IBM’s aim is to encapsulate sophisticated AI models into simple software components that can be easily integrated—much like how developers install common libraries.
Differentiation from Competitors
In terms of differentiation, Bill emphasized the importance of providing robust solutions especially tailored for high-stakes scenarios in regulated industries. This also includes an emphasis on trustworthy AI that encompasses explainability and bias detection, helping companies establish a solid foundation while integrating AI capabilities.
The technology is designed to run anywhere from large-scale cloud environments down to localized, off-grid solutions like speech recognition for point-of-sale systems in restaurants.
Future Plans for Accessibility
Initially, the launch will be high-touch, with select partnerships to gather feedback. However, looking ahead to the future, IBM plans to open-source parts of this technology and enable broader digital self-service access for developers. This approach aligns with IBM’s commitment to making AI accessible to developers at all levels and fostering an environment of innovation.
Bill concluded the discussion by highlighting the importance of developer engagement in adopting these emerging technologies and expressed excitement about collaborating with RedMonk to reach this audience.
Keywords
- IBM
- AI Technologies
- Natural Language Processing
- Speech Recognition
- Foundational AI
- Embeddable AI
- Hybrid Cloud
- Trustworthy AI
- Developer Experience
- Open Source
FAQ
Q1: What is IBM's recent focus in AI technologies?
A1: IBM is focusing on foundational AI technologies, specifically natural language processing (NLP) and speech recognition, to make these capabilities more accessible to software companies.
Q2: How does IBM's AI strategy differ from competitors like Hugging Face?
A2: While Hugging Face provides excellent support for developers, IBM emphasizes high-stakes scenarios in regulated industries and integrates trustworthy AI practices, making it suitable for enterprises that need reliable solutions.
Q3: What are the use cases for IBM's NLP and speech technology?
A3: The technologies can enhance various applications, from customer service automation to data processing and analytics within software products.
Q4: When will developers be able to access these AI technologies?
A4: Initially, IBM will work closely with select design partners, but plans for broader accessibility through open-source options and digital self-service capabilities are set for next year.
Q5: How will IBM ensure the trustworthiness of its AI solutions?
A5: IBM aims to include features such as explainability and bias detection in its AI technologies to support compliance and ethical use in sensitive applications.
