Back to blogs

The top skills for the GenAI era?

Divina Paredes

About 25 days ago By Divina Paredes

Cover Photo (6)

​With the rise of generative AI, businesses are increasing their focus on team members responsible for developing and implementing AI systems: data scientists.

“Our discipline is going to be about radical collaboration with people who built tools that you don’t know exactly all the code that went into that tool… with your teammates who are using complex tools for complex processes,” says Cassie Kozyrkov, Data Scientific CEO and Google's first Chief Decision Scientist.

“Our profession is not about the particular tools we use, our profession is about something more,” says Kozyrkov. “It is about how we think, it is about how we solve problems and there is a lot of toil that folks who are not part of our profession don’t even understand… They think intelligence happens, magic happens, it’s all instant insights.”

Soft skills: ‘I hate that term’

She says questions arise on how AI is going to change the profession. “Now, the tools are easier and the interfaces are getting easier… What is happening to our professions if it is getting democratised, is everybody going to be in it?”

“When we look at generative AI… we’re not taking away the essence of what programming was this whole time. We might be personalising it a little bit, but still the control, the ability to think, speak, and communicate precisely is the core skill that we are dealing with.”

She cites the case of prompt engineering. “Prompt engineering is the same skill I hope every leader has if they are worth their salt. It is giving good instructions. It’s the same skill that a great programmer has. Giving great instructions that the machine can understand so that what I want is what I get. That is prompt engineering. And that is about precise thinking.

“Imprecise communication gives you bad outcomes. So that’s going to be on you. That doesn’t change,” says Kozyrkov. “But as the interfaces get better, as the tools get better, what happens as we start to highlight what makes us data professionals in the first place? And those skills are in danger of being called ‘soft skills’.”

“I hate this term,” she says, of the latter. “You know what soft skills are? They are the skills that are hardest to automate,” says Kozyrkov, who spoke on leveraging AI for better decisions, at the recent SAS Innovate.

“As you take away the fiddly stuff, as the punch cards go away, as the tools become more awesome, it is on you to figure out what to do with them.”

What’s next?

University of Auckland Business School professor Leo Paas says data scientists rely on a mix of skills. “Next to understanding coding and AI analytics, data scientists communicate with managers about strategic and tactical goals.”

“They work with IT on, for example, data quality, model deployment and any potential caveats. Ideally, others who will be involved in the model deployment also are consulted on topics such as law, ethics and upskilling staff to work with the model output. In an evolution of the role, data scientists with prompt engineering skills can now also use generative AI for further advice.”

Data scientists still require a deep understanding of the business context in which they operate, he stresses.

“When speaking with managers they need to have the ability to translate business challenges into data and analytics solutions. Generative AI is based on general rules and averages, while a company’s strategy is influenced by contextual specifics and their dynamics. Conducting analyses, coding and interpreting the output of analyses in terms of the business context therefore require the nuance of a data scientist.”

Looking ahead, Paas says generative AI will first assist in the more technical tasks of data scientists, such as coding and analytics. “Data scientists will increasingly collaborate with generative AI adding their own creativity and intelligence to maintain contribution. Such input about, for example, the business context and its dynamics will continuously improve generative AI models,” he states.

“As generative AI evolves, it can take on less technical tasks, sometimes called soft skills,” says Paas. These include leading meetings about data and analytics as an automated project manager. Or, managing deployment of a model by being a chatbot for the employees who are involved using scores that have been calculated by a newly-developed analytical model.

“Data scientists and others will need to continuously upskill,” he concludes. “Which jobs become relevant as time progresses is difficult to predict and relies on the direction in which the evolutionary progress proceeds.”

Divina Paredes is a New Zealand-based writer interested in #ICTTrends #Tech4Good #DigitalWorkplace #Data4Good #Sustainability #CivilSociety #SpecialNeedsCommunity #SocialEnterprises

Reach her on X (formerly Twitter): @divinap