Generative AI is moving fast. If you don’t stop and look around, you could miss the next game-changing tool for your company. Don’t be that IT leader.
On its face, March announcements from Morgan Stanley and PwC that they were using GPT-4 generative AI software may not have been surprising.
After all, these companies were no strangers to using AI solutions. What was so special about using an emerging AI technology from startup OpenAI to help employees better serve customers?
That perspective undercuts the broader message: That leading global enterprises had planted their stakes in the ground for the new battlefield in AI innovation.
In the past few months, much more of the world has learned that generative AI technologies—which serve up text, images and even software code from basic and layered queries alike—can augment how people work and how companies compete.
More than that, it signaled that generative AI has crossed the rubicon from a fun and experimental virtual assistant to serious business tool—even as OpenAI, fellow startups and high-stakes incumbents continue to refine their products.
Gen AI Has Gone Commercial
Make no mistake: There are a lot of moving parts with which you as an IT leader must familiarize yourself. OpenAI helms this nascent market today, but Microsoft (which has invested $13 billion in OpenAI), Google and others are angling for seats at the high-rollers table. Venture capital firms have invested over $1.7 billion in generative AI solutions over the last three years, according to Gartner. Why should this matter to IT leaders?
It’s still early innings for generative AI and while some organizations have had their fun experimenting, it’s time to get serious. Which means you need to carefully pick which players you want to bet on.
Sixty-five percent of 225 U.S. executives polled by KPMG in March said generative AI will have a high or extremely high impact on their organization in the next three to five years—far above every other emerging technology. Yet 60% of those polled said they are still a year or two away from implementing their first generative AI solution. A year or two away.
Let that sink in.
Caution when adopting any emerging tech is understandable but ask yourself: Can you sit on the sidelines knowing your competitors are scrambling to figure out their strategy for leveraging generative AI?
“There is a true first-mover advantage with the pace of generative AI innovation,” said Todd Lohr, KPMG’s U.S. Technology Consulting Leader. “Winning organizations will establish their competitive advantage by taking decisive action now, while ensuring they are taking the proper steps toward mitigating risk and implementing responsible AI.”
Indeed, whether you elect to experiment with technologies from OpenAI, Microsoft, Google, another startup or all of the above, proceed with the proper governance and guardrails—as you would with any emerging technology. These tips can help you along the way.
Your AI Playbook Needs Guardrails and Governance
The Why Matters. Don’t fall into the trap of wielding a new hammer in search of a nail. Identify a business need or goals to achieve that can be propelled by generative AI. This will help you pick the right tools and approaches, from generative AI apps to supporting infrastructure and reference blueprints. Morgan Stanley, for instance, is looking to turbocharge the way its financial advisors query the organization’s wealth of research. It believes ChatGPT provides such a path.
Prioritize Data Curation. How is your data looking? Like all AI, generative AI models lean heavily on the data with which they’re trained so you’ll want to cultivate and preprocess sound training data to achieve desired results. And as you continue experimenting, you’ll continuously test and monitor results for performance and quality. It’s critical for curbing garbage in, garbage out scenarios.
Keep Humans in the Loop. Ethical concerns are valid in all AI implementations, but none more so than with generative AI. For one, it’s prone to errors known as hallucinations; for another it can fall prey to the same biases humans possess because it feeds on information humans created. Enterprises can’t afford to use faulty data, so have humans fact check the results. PwC, for example, has its legal experts review all results generated by ChatGPT as they conduct their research.
Establish Ethical Frameworks. To help the humans in the loop, you’ll need to educate employees about generative AI and establish guidelines to ensure responsible use. Such guidance will warn employees not to incorporate confidential corporate information and IP while using the technology. Also, organizations must think about their specific risk profiles at every step of implementation to help mitigate algorithmic bias, poor results quality and other considerations. Such practices will help ensure compliance with privacy regulations.
Test and Learn Your Way to Innovation. Employees are using these generative AI tools for personal and work tasks—whether you know about it or not. As an IT leader, you must keep pace with the latest techniques and practices; rest assured, your competition is. Companies that figure out how to accommodate employees using generative AI now will be well positioned. Deploy tiger teams to experiment and tackle new projects leveraging different forms of generative AI. Continuous learning and iteration are key for unlocking game-changing innovation. Ultimately, each enterprise has unique requirements so it’s important to tailor these best practices to your specific context.
As you embark on your generative AI journey, consider this: Some experts believe generative AI is facing its Frankenstein moment, where the technology has the potential to disrupt organizations that cannot rein it in. A kinder analogy is that it’s a toddler technology that requires proper care and feeding.
How are you as an IT leader prepared to help your business navigate this uncertain future?
Learn more about our Dell APEX portfolio of as-a-service offerings and how Project Helix helps organizations leverage generative AI to drive emerging use cases and deliver value.
Read the full article here
