While enterprises are keen to leverage generative AI, they have concerns about transparency, accountability, accuracy and explainability. Responsible AI frames the right kind of boundaries. In fact, they would like to have responsible AI solutions.

Indian IT organizations are, therefore, introducing responsible AI solutions. The need is felt for responsible AI all the more when the model errs in its answers — what we call hallucinations.

In traditional computing, the answers are always predictable (the same inputs provide the same outputs). In AI, there are many probabilistic outcomes whenever it tries to predict. AI system works on probabilistic model, rather than deterministic model. It could be a wide range of probabilities. One may not have the answers one is seeking. Therefore, for mission critical systems, the model must use the right kind of parameters, and provides the right answers.

When guardrails are set, temperatures are also set. In AI parlance, temperature refers to a parameter that is used to control the randomness of the output or answers.

In India, Infosys has launched its responsible AI suite called Infosys Topaz. Coforge, another IT firm has launched Quasar responsible AI solution Harmoni.AI has launched Sonata Software.

All these offerings will accelerate adoption of generative AI at the enterprise level.