There is a lot of excitement in corporate America about the new artificial intelligence (AI) technology that can create business reports and computer code with humanlike fluency.
Many companies are experimenting with generative AI, but they have concerns about data confidentiality, the accuracy of AI-generated answers, and potential legal issues.
IBM has announced its campaign to address these concerns by offering indemnification against copyright claims when using its generative AI systems. Additionally, IBM will publish its data sets, which is not commonly done by other commercial providers of generative AI technology.
This announcement shows that IBM is making moves to tackle the market, while the focus has mainly been on chatbot AI technology like OpenAI’s ChatGPT.
Since IBM’s customers are primarily other businesses, it is crucial to assure them that they can use new AI products without legal issues. Other companies, such as OpenAI and Microsoft, have also taken on legal risks to reassure their customers.
IBM’s AI systems are designed for business use, and the training data is curated from various sources including the internet, academic journals, and legal and finance documents. This commitment by IBM is aligned with the direction the generative AI market is heading.
For corporate customers, it is essential to know the data inputs and understand why a certain answer is generated. Putting confidential data into an AI model is seen as high risk, and IBM aims to address this concern by positioning itself as a partner for companies to create their own AI technology by leveraging IBM’s open models.
While other companies focus on embedding generative AI into existing digital tools, IBM aims to help businesses become creators of generative AI technology. The smaller AI models used in business applications require less computing power and are more accurate, making them suitable for customer service, back-office tasks, and code writing assistants.
In these fields, IBM believes there is a defensible return on investment, and the economics work.