Instant AI answers can trivialise human intelligence, warns Royal Observatory
Baku, May 19, AZERTAC
The rise of AI tools that instantly answer questions and complex problems could make humans less intelligent, the Royal Observatory Greenwich has warned, according to BBC.
The Observatory, one of the UK's oldest purpose-built scientific institutions, is known for its contributions to astronomy.
Paddy Rodgers, director of the Royal Museums Greenwich group which oversees it, said its rich history of research showed the power of human knowledge and curiosity - and the need to avoid "complete dependence" on AI.
"A reliance solely on instant answers risks losing the habits of questioning and evaluation that underpin knowledge, expertise and innovation," he said.
Rodgers' remarks come amid an ongoing transformation of the Royal Observatory in a project called First Light.
The project hopes to "seize on the passion of all the astronomers over the last 350 years, and interpret that passion through science," Rodgers told the BBC.
These discoveries, he said, would not have been possible without technological innovation.
But he added they also would not have occurred without asking and pursuing answers to questions ourselves, and encountering unexpected information or results that AI systems might not relay.
According to Rodgers, early astronomers "built a huge amount of data about the heavens which would subsequently be used for things that they had never thought about," he said.
Their work involved doing unnecessary things "a machine would not do", he told the BBC.
"The human beings did, and it ended up being a huge resource that could be used 150 years after they had written it up to help to verify ideas that people were having about what else impacted navigation on Earth."
At the same time, AI has been used to aid scientific discoveries.
In 2024, computer scientist Sir Demis Hassabis shared the Nobel prize for Chemistry for "revolutionary" work on proteins, the building blocks of life.
Sir Demis, chief executive of Google's AI company DeepMind, used AI to predict the structures of almost all known proteins and created a tool called AlphaFold2.
LinkedIn co-founder and venture capitalist Reid Hoffman described AI as a "transformation" of "cognitive excellence".
"Use it as a counter-agent," he recently told the BBC's Radical podcast.
"E.g. 'What's wrong with my idea?' One of the basic things to use AI [for] is 'I think X, are you against it?'"
Academics and students have also shared experiences of research benefits, including using the tech to challenge ideas or work through solutions collaboratively.
A lecturer at Oxford Brookes University told the BBC last June that "when used responsibly, AI tools enable students to direct their attention to the more important parts of learning and improve their self-development."
But they added that to simply "outsource their thinking" to the tech would highlight its limits.
Limits v promise
Generative AI products that can respond to increasingly complex prompts with text, images, video or audio continue to be developed at pace.
Chatbots have evolved from simple assistants into chatty companions, image generators have become dangerously good at making photorealistic content and new advanced models are said to be surfacing decades-old software bugs.
Such advances, praised and scrutinised in equal measure, are still also accompanied by warnings to users of the tech's limitations and dangers of relying on it.
Rodgers said with previous online tools such as Wikipedia, "if you were interested in something you could perhaps go back to a fundamental source and check it... and see whether or not you found something that was reliable".
Such information can be omitted in quick AI responses, he added, meaning "you're getting more and more distanced from relatable or checkable information".
According to Dr Anuschka Schmitt, assistant professor of information systems at the London School of Economics, "the harmful and unintended consequences of technology, including overreliance, are nothing new".
But conversational AI systems capable of performing many tasks, in a human-like way, have "dramatically reduced the barrier for humans to forego cognitive effort and engagement for work, learning, and leisure," she told the BBC.
Dr Schmitt said with studies of so-called "cognitive outsourcing" revealing "how competencies, memory and learning are quickly but negatively impacted by the use of contemporary AI," it was important to consider when and where to use it.
Nevertheless, generative AI tools that present us with information we do not have to find ourselves are on the rise.
AI Overviews have now replaced snippets or lists of links at the top of Google search results, with similar experiments appearing on social platforms like TikTok and X.
Other news in this section
