Generative AI: A Powerful Tool with Significant Risks

Author William Gibson famously observed, "The future is here already; it's just not evenly distributed." This statement rings especially true for GenAI. The "Chat" is out of the bag, so to speak, and it's being used by several hundred million people already. However, awareness of its full potential and appreciation of inherent risks are still catching up.

To explore how businesses can best utilize GenAI, we invited two experts on-air via InsideAnalysis , the only coast-to-coast radio show all about the Information Economy: Bao-Ha Bui from FPT Software, Americas; and Jim Wilt, a fractional CTO and distinguished architect. These experts provided their insights on leveraging this technology to create business advantage.

Understanding GenAI

At its core, Generative AI is a type of predictive engine. These engines are trained on immense amounts of data and learn to "predict" or generate the kind of text or image a user wants based on prompts.?

Bui noted that the adoption rate for GenAI has far outpaced the cycles for previous technology disruptions. Take cloud computing, for example: "It took 15 years for the cloud to be adopted. A few years ago, some companies were still not looking at the cloud. But contrast that with Gen AI, we've got widespread adoptions in two years."

She continued: "I talked to my son. My son is doing a co-op and I asked him, 'So what are you doing over your break time?' He's using Gen AI to optimize their lab work, such as RNA recognitions and RNA pattern matching. So it goes from a corporate level to students."

She then threw down the gauntlet to traditional thinking about GenAI for students: "The professors are trying to prevent the students from using Gen AI for papers. But is that not a bad thing? Because, at the end of the day, it's reciting information that you can get from Google. Just put it in your own words. It's permeating every aspect of our lives nowadays. So we all have to get onto that bandwagon."

Witt then chimed in about similar adoption curves, noting the amazing rise of GenAI. "Adoption of the PC took about ten years, 15 years to get to 50 million users. It took about ten years to 15 years for cloud adoption to really land. And when we look at machine learning and Generative AI, machine learning has been around for ten years at least. There are people that have been working in this space for over 20, 30 years with neural networks and so forth."

He continued: "But the big difference with Generative AI is it's really applicable to the common person. Anybody can get into a prompt situation, learn some prompt engineering, and start getting results that are just massively interesting. And I think that the statistic that hit me that Bao-Ha mentioned is actually that Generative AI has reached a hundred million users in just two months!"

Low-Hanging Fruit

GenAI is not just fascinating; it's very useful for all kinds of business transformation. Its benefits can span many critical areas, including:

• Content Creation:? GenAI can take the drudgery out of creating marketing copy, product descriptions, social media posts, and email newsletters. It frees up your team to focus on high-level strategy and creative execution.
• Product Development: GenAI sparks innovation, providing a stream of new product ideas, designs, and prototypes. It can also automate tasks like market research and competitor analysis, all of which accelerate product development.
• Customer Service: GenAI can fuel chatbots that deftly answer customer questions and resolve issues around the clock. The result? Happier customers and reduced support costs.
• Personalization: GenAI tailors experiences to individual customers by generating content, recommendations, and offers that resonate with their unique needs and preferences.
• Efficiency and Productivity: GenAI automates repetitive tasks across the board, leading to improved efficiency, streamlined processes, and greater productivity throughout the organization.

Getting Started

If you're ready to implement GenAI, Bui and Witt offered some key points to remember. First things first: Identify your use cases! Where can GenAI step in to either improve what you're already doing or handle tasks that are currently bottlenecks? Repetitive marketing tasks? Customer service? Brainstorming sessions?? Pinpoint specific areas where GenAI can make the biggest impact responsibly.

Bui made an excellent suggestion that a first use case should be internal, in other words, not customer-facing. Employees can get a feel for how these engines work by playing around with the technology internally. One critical best practice? Pay attention to details with your prompts, because slight changes can significantly impact results.?

The ancient Greek philosopher Heraclitus offers some perspective here. He's the wise man who once quipped: "No man ever steps in the same river twice, for it's not the same river, and he's not the same man." So it is with GenAI: Because the foundation of this technology is probabilistic mathematics, you will always get a different result when you change your prompt. This difference can be slight or significant, even if the change in your prompt seems inconsequential.?

The probabilistic nature of GenAI is the kernel of its capabilities but also the source of a major risk, sometimes called hallucinations. In short, GenAI makes stuff up, and that's what it was designed to do. Most experts understand this, but the general public and business world still need education to be aware of the possibility that these engines will fabricate information.?

At a recent event in the US Capitol, Dr. Catherine Ordun, VP of AI for Booz Allen, summed it up like this: "Anyone who tells you they can get rid of hallucinations is either lying, or ignorant." She cautions that use cases must be carefully chosen and then output monitored closely to mitigate the potential for unwanted results.?

Even the legendary futurist Ray Kurzwell recently chimed in on this topic via a Joe Rogan podcast, noting that these AI engines don't yet know how to say: "I don't know!" Kurzwell has received 21 honorary doctorates and has been thrice-honored by US Presidents. He says that work in this space is ongoing.?

Organizational Readiness and Human Oversight

Both Bui and Witt agreed that humans in the loop will be crucial for success with AI. Companies must establish governance policies to ensure the quality of GenAI content. Processes should be established for reviewing and editing GenAI content before it goes live.?

Data is the blood of AI. Organizations must also examine the condition of their data practices. Do they have a good understanding, management, and governance of their data assets? If their data is not ready, their AI is not ready. Pushing AI before Data is ready is a warranty for failure.

There are also significant ethical concerns with GenAI, especially black box versions like OpenAI's ChatGPT. The engine's lack of transparency means there will be questions that cannot be answered about exactly how it decides to create original content. A guardrail of human oversight is, therefore, a strong requirement to avoid pitfalls when leveraging this tech.?

But these concerns should not derail your exploration of GenAI. Witt explained: "If you have a policy in your organization not to allow people to use Generative AI, be aware they're using it no matter what. It's on their phone. And I know that we want to protect corporate assets and corporate secrets. Fine, that's great. But I think if you go about the adoption we talked about, you'll find you can ease your way into it and get many of those fears dismantled and a lot of the excitement generated because you're doing it with internal uses."

Witt noted that GenAI is extremely useful as a co-pilot, especially for developers. "I think it's a code companion. Every developer should have it tied to their hip because it will help them find errors they're putting in their code quickly. You don't let it write the code for you outright. What you do is it's your companion. It's going to find your blind spots; it's going to find where you made a mistake."

Bui concluded: "I think that as great as our human brains are, we have limited capacity of how fast we can generate or how much we can retrieve. So I think it's more of a retrieval ability, whereas the machines will take everything it knows and give it back to you. And it reminds you of things that, hey, I knew this, but I didn't remember, and that's where the benefits or the efficiency gains in using Generative AI are so great because it really takes away the need for us to remember everything."