Coding Isn't Dead—But Here’s How It Should Evolve, Says Google DeepMind Scientist

• Stefania Druga works as a research scientist at Google DeepMind and is the founder of the AI education platform called Cognimates.
• Early exposure to AI education assists children in forming positive interactions with technology as they discover how to "collaborate" rather than deceive, she explained to BI.
• Druga indicated that abilities in artificial intelligence and programming will remain pertinent going forward, yet the approach to teaching them requires an update.

Google DeepMind researcher Stefania Druga aims to inspire children to utilize AI for collaboration instead of using it as a shortcut.

And although there has been significant discussion regarding whether "learn to code "is still solid career guidance," she explained to Business Insider that she thinks coding abilities will remain significant—but the way they teach these skills in schools requires an update.

Conversations surrounding AI in education often highlight the possible drawbacks, such as cheating and skill atrophy Druga mentioned that young individuals often heavily rely on artificial intelligence, yet they do not always use it towards achieving their best outcomes.

"As I see it, how they’re currently utilizing it is quite disappointing,” Druga stated. “However, this isn’t their mistake; it’s due to the design of these technologies. Using AI for education like getting assistance with homework, writing essays, or passing tests isn’t the most effective application.”

Druga mentioned that she thinks one issue could be the type of tasks students are required to undertake.

She suggested we need to overhaul the entire system. Primarily, if artificial intelligence can pass a test, it indicates that the test itself may be flawed," she explained. "Furthermore, concerning essay generation—if you rely on a tool that provides only predetermined data without allowing any interaction or feedback, this leads to excessive dependence and hinders critical analysis skills.

Druga mentioned that she initially recognized the necessity of AI education almost ten years ago — well before the general public was familiar with terms like artificial intelligence. LLM was.

"Alexa would be the go-to source for half of U.S. households, with individuals turning to this voice assistant before even consulting their parents," Druga stated to Business Insider.

Druga — also shares responsibility for Scratch , a drag-and-drop programming language used to teach kids how to code — created Cognimates As part of her graduate research at MIT, this initiative promotes AI education by involving children in captivating activities such as developing games and coding robots, alongside training them to work with artificial intelligence models.

The aim is to offer children an environment where they can safely explore technologies that will probably play a major role in their lives for years to come.

“It’s essentially building an environment where children can dive into the world of science through play, as they develop theories such as ‘Here’s why Alexa reacts this way.’ They also get to test these ideas rapidly,” Draga explained. “Similarly, with the advent of Gemini, ChatGPT, and big language models, how do we empower younger generations to design their own versions of GPT?”

Druga mentioned that she created her own platform with a more Socratic approach—it encourages users by asking questions that guide them towards the correct path instead of providing direct answers. Additionally, she noted that the type of challenges kids tackle often make them feel quite accomplished and thus very attached to their creations.

Their attachment to their project is quite strong; it forms a significant portion of who they are," explained Draga. "They feel possessive about it, thinking ‘This is my creation, and I don’t want an AI taking over.’ However, whenever they hit a roadblock, they wish they had assistance with debugging or locating specific blocks, as well as navigating through the platform itself.

Drugai suggests that fostering both critical thinking skills and AI literacy from an early age, starting when children first learn to talk, requires a dual strategy. In educational contexts, AI tools ought to facilitate collaborative creation instead of merely aiding students to find answers swiftly. Additionally, tasks should be crafted with more open-ended parameters rather than being straightforward and definitive.

It’s not fair to expect them to consistently make the correct decision, as I believe that’s asking too much," Druga stated. "If you had tedious homework and an instrument capable of completing it instead, why would you refrain from using it? I don’t fault them for it.

She mentioned, 'We must alter our methods of teaching and evaluating.' Additionally, she stressed the importance of modifying the design of these tools to accommodate the autonomy and creative potential of young individuals.'

Coding isn't obsolete — however, coding education requires an update.

Druga mentioned that Cognimates is equipping kids with abilities that will be relevant in a workforce that hasn’t materialized yet. Although she often encounters the claim that “coding is obsolete” (a notion she disagrees with), she still considers foundational knowledge, including an awareness of artificial intelligence, to be highly valuable.

Druga pointed out that one issue with computer science and computing education for many years has been their emphasis on the job market and readying students for careers in technology. It was often suggested that obtaining a CS degree guaranteed an easy job without much concern over future challenges.

Druga pointed out that this statement is no longer accurate, considering the current instability in the employment sector. sweeping layoffs and concerns about AI replacing human workers , but she thinks that goal was the "incorrect type of objective" to start with.

“What we’re observing with AI currently – particularly within large language models and subsequent architectural developments – is that this field is evolving at such an accelerated pace that if your primary approach to training individuals revolves around equipping them for particular technologies or roles, it will quickly become outdated,” she explained.

Druga suggested that preparing for a career in technology ideally entails developing transferable skills, with an emphasis on teaching individuals how to navigate uncertainty. She anticipates that the sole unchanging factor will be "constant change, and quick change." This perspective highlights her views on the industry’s dynamic nature. The CEO of Google DeepMind, Demis Hassabis, Shares. Although the conventional tech career trajectory might not be feasible when today’s Cognimates-using children graduate and start looking for jobs, Draga thinks it could potentially become irrelevant as well.

“I can say firsthand that I know around 20 individuals who were 19 years old and managed to go from a prototype stage to achieving full financial sustainability within a few months,” she stated. “These folks didn’t require venture capital funding or work in technology. Instead, they focused on building solutions directly. Once they identified an issue, developed their answer, launched it, and started earning revenue; they became profitable. That’s really something when we discuss the future of employment, isn’t it?”

Drugai thinks it doesn't matter what path children choose regarding their AI abilities; the crucial part is that they cultivate these skills from the start.

"The aim is to ensure that everybody believes this initiative is meant for them and doesn’t feel daunted or thinks they must possess extensive prior knowledge just to begin, as everything evolves rapidly,” she explained to BI. “I believe the frameworks for AI engineers or scientists are truly reshaping our perspective on workforce education and how traditional pipelines function.”

If you liked this tale, make sure to follow Business Insider on MSN.