Participants agreed that the latest set of advancements in GenAI posed unprecedented and unknown implications for society, and therefore, the use of such a transformative technology demanded careful calculation and supervision.

Participants suggested two distinct approaches for GenAI adoption depending on the context of use of the technology.

The first approach (deny all, allow some) should apply to the use of a GenAI solution by an enterprise or organisation to meet external customer demands. In this approach, until robust risk mitigation frameworks are figured out to ensure responsible development and use of GenAI for public benefit, only GenAI solutions that are truly compelling should be rolled out for public adoption as an exceptional response to a given problem.

The second approach (allow all, deny some) should apply to the use of a GenAI solution by an enterprise or organisation to improve internal workforce productivity. In this approach, the enterprise or organisation, by default, should allow deployment of all GenAI solutions, but in a restricted environment, to improve worker productivity, while making an exception for solutions that pose significant inherent risk to workers.

Participants echoed the need for businesses to be motivated by a strong, socially beneficial purpose when scoping GenAI for integration into their workflows.

For instance, some participants highlighted the technology’s potential to improve insurance penetration in the country: if insurance policies could be communicated in local languages such that the need for an insurance mediator was obviated, it would reduce the cost of insurance premiums, thereby making insurance protection more affordable. In addition, socially beneficial use of the technology were underlined in healthcare (helping drug innovation), agriculture (helping farmers with crop-related queries), and government administration (making government policies more accessible to the public).

Participants underlined the importance of risk anticipation at the stage of model inception to guide model fine-tuning and risk governance for responsible and safe model deployment.

Participants expressed concerns over the current regulatory uncertainty hovering over GenAI development and use, making compliance in the industry a big challenge.