Published on November 1, 2023, 8:01 am

TLDR: One year after the introduction of ChatGPT, IT leaders are cautiously exploring generative AI but are hesitant to fully embrace it due to the technology currently falling short of expectations. However, there is confidence that generative AI will become indispensable once it matures and can automate mundane tasks. Some predict that generative AI will be more disruptive than the iPhone for consumers and have a greater impact on workforce productivity than Microsoft. While AI itself is not new, generative AI is taking traditional AI to the next level by enabling more conversational interactions and providing deeper insights. However, there are obstacles to widespread adoption, such as a lack of skills among personnel and confusion among vendors regarding integration. Organizations need to decide whether to build or buy generative AI solutions and develop in-house expertise for success. Failure to embrace this technology may lead to becoming obsolete in rapidly evolving industries.

A year after the introduction of ChatGPT, IT leaders are still hesitant to fully embrace generative AI, opting for cautious exploration instead. While many CIOs recognize the potential benefits of generative AI, they find that the technology currently available on the market falls short of expectations.

Yves Caseau, the global CIO of Michelin, has been experimenting with both GitHub Copilot and ChatGPT for over six months. He acknowledges the rapid evolution of generative AI but believes it is still in its early stages and is more like a toolbox than a fully matured solution.

Despite some skepticism surrounding these large language models (LLMs) and associated tools, there is confidence that generative AI will eventually become indispensable. Caseau believes that once it matures, generative AI will automate mundane tasks, allowing organizations to focus on more innovative endeavors.

There are technology leaders who go even further by stating that generative AI will be more disruptive than Apple’s iPhone for consumers and surpass Microsoft’s impact on workforce productivity. Patrick Thompson, former chief information and digital transformation officer of Albemarle, predicts that generative AI will be the most disruptive technology in our lifetime.

While generative AI is relatively new, AI itself is not. Predictive maintenance using artificial intelligence has been one of the earliest use cases in many companies. By training algorithms on data collected by sensors, companies can identify indicators that may lead to failures and prevent manufacturing outages.

However, one limitation of predictive maintenance is that rare events are often underrepresented in training data. To address this gap in knowledge, many companies supplement real data with synthetic data.

AI is also being applied in various other ways within enterprises to improve supply chain efficiency, facilitate customer interactions, and assist employees with office tasks. Albemarle used AI as a virtual assistant during the pandemic lockdowns to support self-servicing for 7,000 employees working remotely.

Generative AI is now taking traditional AI to the next level, particularly in applications like predictive maintenance. Interactions become more conversational, enabling operators to ask questions and gain deeper insights about the state of equipment. It can also curate industry data to train traditional algorithms and provide agile results.

Generative AI offers an entry point for sectors that have yet to fully embrace traditional AI. Companies in finance, for example, are leveraging their existing data platforms along with analytical tools to experiment with generative AI technology. It is being used to parse publicly available data on markets and companies, assisting in investment decision-making.

However, there are several obstacles that need to be overcome before generative AI can be widely adopted by enterprises. One major challenge is the lack of skills among both in-house personnel and vendors specializing in traditional applications.

Organizations must make a build versus buy decision when it comes to generative AI implementation. Buying cloud-based models may be cost-effective and expedite adoption initially, but building in-house solutions allows for customization according to specific industry needs or running inference on edge devices.

Currently, very few enterprises have the expertise required to develop or fine-tune AI models. To maximize the benefits of purchased models, companies need to curate their enterprise data for training purposes and learn how to ask questions effectively during the inference phase.

Another issue affecting the adoption of generative AI is confusion among traditional application vendors regarding integration with their existing products. Partnership with dedicated generative AI developers is crucial for delivering virtual assistants that unlock the value of enterprise systems while balancing security and privacy concerns.

Although many organizations experimenting with generative AI are large enough to afford such endeavors, this technology is not limited to big enterprises alone. With proper governance, security measures, and data ingestion strategies in place, small companies can also leverage generative AI to scale their operations efficiently.

The lack of widespread adoption of generative AI may create a digital divide between digitally savvy companies and those lagging behind. Regardless of whether organizations choose to build or buy generative AI solutions, developing in-house expertise is crucial for long-term success.

In conclusion, while there is still room for improvement and challenges to overcome, generative AI has the potential to revolutionize industries and increase shareholder value. Companies that fail to embrace generative AI risk becoming obsolete in the face of rapidly evolving technologies.


Comments are closed.