Google Deepmind’s researcher Stefania Druga wants to encourage children to use AI to “co-create” rather than cheating.
And although there has been a lot of debate on the question of whether “learning to code” is always good career advice, she told Business Insider that she thought that coding skills will continue to be precious – but how they are taught at school.
Discussions on AI in education are often marked by potential negatives, including cheating and skills atrophy. Young people said Druga, are often AI users – but they don’t always draw it for the best ends.
“The way they use it right now, in my opinion, is really disappointing,” said Druga. “But it is not their fault. This is how these technologies were designed. It is not the best thing we can do with AI for learning, to make it resolve our homework, or write our tests, or help us to take a test.”
Druga said she thought that part of the problem is the type of work that students should finish.
“We should change the entire framework. First of all, if an AI can solve a test, this is the bad test,” she said. “And then secondly, for the generation of tests-if you have a tool that always gives you pre-cooked information, and that there is no back and forth, this is when you have problems of renown and lack of critical thinking.”
Druga said that she had first carried out the need for AI education almost a decade – long before the average person knows what an LLM was.
“Half of households in the United States would have a voice assistant, and people would first ask Alexa to ask their parents,” Druga told BI.
DRUGA – Also partially responsible for Scratch, a drag and drop programming language used to teach children to code – has created cognization as part of the thesis of his MIT control. The program encourages literacy of AI by engaging children in projects that precipitate their interest, including the construction of games and programming robots, as well as learning to form AI models.
The idea is to provide children with a platform to safely experiment with technology that will probably be part of their lives in the foreseeable future.
“It is a kind of creation of this sandbox or this playground so that children engage in the scientific process, because they formulate hypotheses like:” This is why Alexa responds in this way. “And then they have a way to test this hypothesis very quickly,” said Druga. “And the same thing would apply to the era of Gemini, Chatgpt and models of great language. How to allow young people to create their own GPTs?”
Druga said that she had designed her own platform to be more Socratic in nature – she poses questions to lead users along the right track, without nourishing the answer. And the kind of problem that children are invited to solve, she added, mean that they are often proud and therefore possessive, on their work.
“They are very attached to their project. It is a large part of their identity,” said Druga. “So they are like” it’s my project. I don’t want AI to do it for me. “But when they are stuck, they would love to have someone to help them debug, or to help them find the right block, or to help them even navigate the platform.”
According to Druga, the approach to preserving critical thinking while also ensuring that children begin to develop literacy of AI “as soon as they can speak,” should be two -parts: when used in educational contexts, AI models should support “co -creation” rather than helping students reach a finite response as quickly as possible – and the assignments could be designed to be less cut.
“The burden should not be on them to always make the right choice, because I think it’s too much to ask,” said Druga. “If you were given boring homework and a tool that can do it for you, why wouldn’t you use it? I don’t hold it against them.”
“We have to change the way we teach and assess,” she added. “But we must also change the way we conceive these tools in order to make room for the Young Agency, the creativity of young people.”
The coding is not dead – but the education of the coding must be refreshed
Druga said that Cognimats teaches children skills that will serve them on a job market that does not yet exist. And although she frequently hears that “coding is dead” (what she does not think is true) – she always considers education in fundamental principles, which now include an understanding of AI, as extremely useful.
“I think that a problem with CS education and computer education for the longest was that it would focus on the market, and in a way the preparation of young people to find a job in technology,” said Druga. “It was this promise, like:” Oh, if you have a CS diploma, you will have a comfortable job and not have to worry about anything. “”
This is no longer true, said Druga, given the uncertainty in the labor market due to scans and concerns about the replacement of human workers, but she thinks that it was the “bad type of objective” to start.
“What we see with AI at the moment – in models of great language and other architectures that come after major language models – is that technology changes so quickly that if your full value proposal for the way you train people is to prepare them for specific batteries or jobs, this will become obsolete very quickly,” she said.
Ideally, Druga said that preparation for work in technology involves transferable skills, in particular teaching people to adapt to ambiguity, because the only constant it provides is “change and rapid change” – an opinion that the CEO of Google Deepmind Demis Hassabis Shares. Although the traditional technological career journey is not viable when children currently use cognitive are ready to enter the job market, Druga thinks that this may also not be necessary.
“I know personally, at least, like 20 people who are 19 years old and who went from a prototype to a fully sustainable product financially in a few months,” she said. “They don’t even need VC money or a job in technology. They are building themselves. They identify a problem, they build a solution, they launch it and they are paid for it. It’s profitable. So talk about the future of jobs, right?”
Regardless of what children end up doing with their skills in AI – Druga believes that what is important is that they develop them in the first place.
“The objective is to make sure everyone has the impression that it is for me, and they do not feel intimidated, or they do not have the impression:” Oh, I need to have all this basic knowledge to start even “, because things change so quickly,” she told Bi. “I think that IA engineering executives or AI scientists are really contesting how we thought about the education of the labor market, and in a way the pipeline before.”
businessinsider
