Adoption Of Generative AI: What Should Enterprises Consider?

GenerativeAI
GenerativeAI

ChatGPT and Dalle-E are the talks of the town as the new shiny object that could potentially disrupt Google’s hegemony. The hype cycle, as usual, is high. AI dominated conversations around tech at this year’s World Economic Forum in Davos, Switzerland. Even at this year’s CES trade show, hundreds of exhibitors were listed under the show’s artificial intelligence category—double those categorized as metaverse, cryptocurrency and blockchain combined.

There is no denying the strength that generative AI is demonstrating. We must, however, take it with a pinch of salt. One of the core concerns is that it is a massive model with 175 billion parameters. Therefore, it’s uncertain what the budget impact would be to use it. Some new research from DeepMind aims to reduce the size of these models and, subsequently, the costs, which will likely impact enterprise-level adoption of these tools significantly.

From generative AI’s implicit biases to the fact that its responses depend on the language patterns it has learned rather than any world observations, there are several reasons to be skeptical of the current state of generative AI. To ensure this new technology is approached with caution, it is essential to thoroughly evaluate how it can be incorporated into your company’s processes.

Enterprises that are deliberating on adopting AI into their businesses should be cautious in adopting generative technologies. Ideally, one would want to use vendor APIs, but there are issues related to cost and customer data privacy. Therefore, the best approach, for now, is to start with publicly available data such as documentation or marketing collaterals. This can be used to experiment with vendor services to front documentation with a chatbot, introduce a writing assistant for marketing, etc.

Source