With the shifting landscape of the analytics and data science world, much of it impacted by developments in AI, it has become increasingly important to gauge the perspectives of the industry. These 10 data science thought leaders weigh in on the changes taking place associated with AI, and they share their perspectives on a range of topics from what’s on their desk this year, to what needs improvement in the field.
Download the Artificial Intelligence/Analytics & Data Science Perspectives Report Now
Learning About LLMs
With artificial intelligence technology and LLM systems becoming increasingly developed, strategized upon and implemented, it has become more important to gauge just how this new tool is impacting disciplines such as insights, innovation, analytics and data science.
Data science leaders from a broad range of industries look at their respective positions through a lens of change, disruption and ultimately transformation taking place in the business world, and how it affects the analytics discipline.
Setting strategy seems to be one of the first priorities at this critical juncture, notes June Dershewitz, Co-Founder, InvestInData. There may still be much experimentation and testing of AI and LLM systems, and everything is being explored from in-house options to outsourcing methods using specific vendors who have developed their own operations.
If 2023 was the year of hype and experimentation, 2024 seems to indicate a more practical way of thinking about these high-tech developments.
“When everyone began evaluating LLMs in 2023, we could have been in the position to set some goals in mid-year to see what we’re going to achieve by the end of 2023. And so those goals just hypothetically might look like we’ve evaluated some opportunity in these key areas. The next logical step is to run proof of concepts to see if there’s anything there. Successful proof of concepts would allow leaders to find out where the nuggets of actual interesting buyable things are. In 2024, it would be about pursuing those ideas, maybe as a full rollout,” says Dershewitz.
The pace of investment into LLM’s is also accelerating, notes Bob Bress, Vice President and Head of Data Science at Freewheel, a Comcast Company. “We are seeing significant investment into the development of LLMs and LLM-based technology solutions,” he says. “This investment will accelerate the capabilities of these models for industry-specific use-cases. Companies that lean into new LLM capabilities have an opportunity to find themselves ahead of the curve in terms of creating business efficiencies and enhancements for their products and services.”
Sharpening the Skillset
Connected to exploring these AI and LLM systems, there is an equally corresponding urgency in terms of education and training—whether that means hiring someone new, upskilling or reskilling team members already in place, or a mix of both methods.
If AI and LLM have heightened one aspect of data science, it is the theme of data accessibility and data governance. It seems that no longer can the data scientist be the only “gatekeeper,” so to speak. Data is meant to be accessed by all and leveraged by all parties, for a more interdisciplinary and cross collaborative approach to the field to help the company achieve its overall goals.
For Michelle Ballen-Griffin, Head of Data Analytics, Future, the future is about taking that data science foundation to the next level, creating strong partnerships, and influencing the business. But everyone in data science should prepare to level up.
“As for what must be improved within the discipline, I think it’s the skill set. It’s leveling up,” says Ballen-Griffin. “You need to be technical, strategic, and proactive. What needs to be improved within this plan is giving people opportunities to improve, understanding what an effective data partnership looks like and how to mirror that more. How do I just continue to advance that in a way that makes my discipline more effective for the business?”
Others see that the fundamentals of data science might be the same, but that skill sets are becoming more specialized—at least, until AI becomes more democratized, which other executives also point out might be the case in the not-so-distant future.
“The longer-term trend is really one of specialization,” observes Michael Bagalman, Vice President, Business Intelligence and Data Science, STARZ. “Data science is moving from its youth into its middle age at this point. We are seeing people that are starting to subspecialize seriously within data science. I think that has tremendous implications both for the training of future data scientists as well as their career paths.”
He adds, however, “As you get up to levels of management and especially levels of leadership, you need broad exposure. No one can be deeply expert in every aspect of data science. But at a leadership level, you need to have at least a medium depth across a wide variety.”
The Development Pace Accelerates
To be sure, whether your focus is on a foundational element or a specialized segment, data science remains a hot commodity as companies ramp up their investments into LLMs, and as the pace of innovation within the analytics space accelerates.
“As the pace of innovation within the analytics space grows, we can expect that we will need to adapt and learn at a faster pace than before,” says Bress. “There will always be uncertainty, but we do know technology will progress at an accelerating rate. Technology progression in all fields will be driven by advances in AI, and the companies that adjust quickly and embrace the change will be best positioned to grow and succeed.”
With uncertainty in the market environment a constant, Google’s Chief Strategist, Neil Hoyne, says now is the time to invest in data science, manage risk and move the business forward. It is an exciting time with events moving at a dynamic pace.
“I’m pretty excited about LLMs and how they’re starting to capture the attention of real AI applications in businesses. These tools are making AI tangible, accessible, and definitely exciting,” says Hoyne.
He adds, “The area that also has my interest is the impact of these AI tools on existing business processes. Everyone’s talking about the big picture: how AI is going to revolutionize this industry or disrupt that one. Great areas to discuss. But there’s a huge gap in understanding how AI will be integrated into businesses themselves. Are entire departments going to be overhauled? How will non-AI teams work with AI-driven ones? What functions are actually slowing this transformation down, and how can we revamp them to better support AI integration?”
Future Investments
Just how much investment, and how much the pace will accelerate, is anyone’s guess in terms of AI integration into corporate enterprise.
For Bill Shander, Information Designer/Educator, Beehive Media, data science is still on solid ground, especially in terms of the human element that perhaps an AI cannot replicate, such as consulting, data storytelling and influencing the business.
“These LLMs are amazing, as far as what they can do from a standpoint of answering questions, and so on,” says Shander. “But when it comes to knowing what’s going on in your business, are the tools out there implementing AI in a way that’s really managing the data analytics and the understanding—especially the communications of that to people so they can make their own sound data decisions? We’re still a far distance off from that.”
For Meltem Ballan, Chief Executive Officer and Co-founder, Concrete Engine, corporations are still trying to figure out the best investment approach—and that could take a while, especially if AI is considered part of a digital transformation effort. As some companies undergo layoffs and cost cutting moves, it impacts their ability to invest in technology.
“Companies are not quite ready to invest in LLM and generative AI to see the full capacity and impacts on revenue. The market confusion and hype slowed down initiatives such as completing their digital transformation. Data and platform readiness is work for 2024,” advises Ballan.
The Pillars of ‘Responsible’ AI
While the possibility of AI is endless, questions remain regarding the fairness, transparency, and privacy of this new data science technology.
For Serena Huang, Founder of Data With Serena, the pillars of responsible AI are coming to the forefront as the conversation continues around AI and its components.
“Responsible AI is coming up a lot as a topic in my conversations with leaders across industries,” says Huang. “There are multiple pillars of Responsible AI: fairness; how might an AI system allocate opportunities, resources, or information in ways that are fair to the humans who use it? There is also the matter of transparency; how might people misunderstand, misuse, or incorrectly estimate the capabilities of the system? There is also the question of privacy; responsible AI systems should be developed with values such as anonymity and confidentiality.”
Just what the human element and the AI element can do together remains to be seen, although the promise and potential of this new tool is boundless.
“Ultimately, the possibilities of LLMs are endless. But there is an interplay at work here of what can be automated by technology and what should remain the element of the human employee. I remain fascinated and hopeful as a data practitioner,” says Huang.
Download the Artificial Intelligence/Analytics & Data Science Perspectives Report Now
Video courtesy of IBM Technology
