Leveraging generative AI in your business: mastering the "sweet spot" of knowledge and risk
A practical frame for AI leaders who explore business scenarios for generative AI
ChatGPT and generative AI have been taking the world by storm since the end of last year. The ups and downs, the risks and opportunities, the examples of helping and hindering, and the big bets and losses — it’s been an eventful start to the new year to say the least. Business leaders’ continue to push their AI teams to find the next idea where generative AI can make or save the company money. But despite the excitement and hype — yes, even because of it — AI leaders need to advise their leadership and manage expectations.
» Watch the latest episodes of “What’s the BUZZ?” on YouTube or listen to it wherever you get your podcasts. «
The strengths and weaknesses of large language models
One of the most prominent examples of generative AI are large language models (LLM). These AI models have been trained on vast amounts of data. They are extremely good at generating text output that is nearly indistinguishable from how humans write. But the more people use this technology and tools like ChatGPT, the more their limitations surface as well. After all, LLMs are very good at predicting the next word in a sentence. And that’s just about it. They don’t have knowledge. They don’t retain it. They cannot tell fact from falsehood. And that is important to remember as the output reads seemingly human.
We are seeing more of these limitations reach the headlines as well. For example, New York Times columnist, Kevin Roose’s, encounter with Bing AI in which the chatbot gives increasingly creepy answers, urges him to leave his wife, and “just wants to be loved 😘.” These uncontrolled scenarios and responses can pose significant reputational, financial, and legal risks for businesses that use the technology. But AI leaders can help their business counterparts determine and evaluate opportunities for generative AI in their business and find the sweet spot.
The “sweet spot” for generative AI
Generative AI tools like ChatGPT can be a significant productivity booster for professionals in business. What has taken minutes or hours to create, can now be accomplished in seconds. However, professionals shouldn’t trust the generated output blindly. The previously mentioned business risks of AI creating factually inaccurate or unwanted output are all too real. Just because it reads as if it were factually accurate, does not mean that it automatically is. This has the potential to put the “post-factual world” into overdrive. Users who have (at least some level of) knowledge about the topic they’re using Generative AI for can determine the accuracy of the generated output more easily.
The sweet spot for using generative AI in business is currently at the intersection of supporting individuals who have pre-existing knowledge about the information they seek (+) to turn AI-generated recommendations into decisions with a low risk to the business (-).
Examples of such “sweet spot” scenarios include:
Propose meeting agenda
Recommend interview questions
Act like a management consultant
Incorporate keywords in an introduction
Tighten or summarize long-form text
While this is an initial attempt at structuring low-risk use cases in business to explore, it is only a snapshot in time. Both, the technology as well as its users will evolve over the coming months as generative AI becomes more robust and its users become more proficient in using it.
For now: Using AI to optimize what you already know — is that too conservative?
» Watch the latest episodes of “What’s the BUZZ?” on YouTube or listen to it wherever you get your podcasts. «
What’s next?
Appearances
March 28 - Join me for a virtual conference hosted by SwissCognitive together with 20 Global AI Ambassadors and experts on Redefining Business Performance with Generative AI.
Join us for the upcoming episodes of “What’s the BUZZ?”:
March 14 - Tom H. Davenport, Professor and Author, and I will talk about Being All-in On Generative AI — and what that entails. Join 150+ professionals live to have your questions answered on the show.
Follow me on LinkedIn for daily posts about how you can set up & scale your AI program in the enterprise. Activate notifications (🔔) and never miss an update.
Together, let’s turn hype into outcome. 👍🏻
—Andreas