By: Aditya Gupta
The emergence and exponential evolution of artificial intelligence across various sectors, particularly business, has ignited discussions about its implications within entrepreneurship.
It is beyond clear the practical utilization of AI has reduced the barriers of entry for new business ventures, enabling entrepreneurs to automate redundant tasks and decrease their dependency on employees or contractors. Among these tasks include automating newsletters, predicting ad performance, improving customer service through chatbots, following up with leads, or making accounting easier. Frankly, entrepreneurs who fail to effectively utilize AI tools to accomplish the aforementioned tasks will be squandering resources and time, resulting in failure to maximize operations and/or competitor success.
With the evolutionary nature of the development of AI technology, a new facet of AI moves beyond automation and the completion of repetitive tasks. Companies like IBM1 have begun rolling out advanced generative AI models coined as Consulting AI Services, which utilize consumer and company data to generate optimal business decisions for customers in order to forecast improved KPIs (key performance indicators) and trends. This is amazing. Now, employees no longer have to dig through industry research, sales trends, and marketing metrics to craft business decisions. Step aside and allow a machine to do it for you.
So why the title? Well, there are key factors to consider about the nature of big business in regards to the materialization of advanced generative AI models. It is no secret that large tech corporations clandestinely hold onto new artificial intelligence technologies they develop. The public only finds out about the technology months after, at best, and it is often years before the technology is sold to other businesses and typical consumers receive it last. This allows large corporations, particularly in the tech space, to always hold a competitive advantage over those entering the market, as they possess the resources and capabilities to operate large-scaled AI development projects (and the ability to incur losses). And even if we optimistically assume these companies do share their innovations or argue this is simply the rewards of the capitalistic system, another unavoidable problem arises. Data.
As Stanford’s Institute for Human-Centered Artificial Intelligence concluded, “AI Models Are Only as Good as Their Data Pipeline.”2 Are the datasets possessed by an SME tech company comparable to those possessed by the likes of Salesforce or SAP? Utilizing an AI model with limited data is almost objectively worse than asking a human to interpret the same data. And if we truly believe, as I do, that these data-driven AI models effectively generate business decisions which tangibly lead to better business outcomes and an increase in KPIs, the more high-quality data it is trained on, the better a business will do. Smaller tech companies simply do not have the data that a company like Salesforce does. Salesforce has access to the accounts, leads, opportunities, etc, of over 150,000 customers. This level of accumulation of data renders AI models like IBM’s as bulletproof business strategy developers. How are small businesses meant to keep up with this?
Adding the enormous access to data of large corporations with the strong likelihood that most advanced artificial intelligence models are withheld by these same corporations, the path for a small business to compete with these companies seems a near impossible task. However, all hope is not lost for small businesses. Due to the aforementioned ever-evolving nature of AI technology, it is certain that solutions to such predicaments will arise.
References
1. “Artificial Intelligence (AI) Services & Consulting.” n.d. www.ibm.com. https://www.ibm.com/consulting/artificial-intelligence.
2. Miller, Katharine. 2022. “Data-Centric AI: AI Models Are Only as Good as Their Data Pipeline.” Stanford HAI. January 25, 2022. https://hai.stanford.edu/news/data-centric-ai-ai-models-are-only-good-their-data-pipeline.
Comments