Science & Tech
Daily Show’s Jordan Klepper trolls Melania Trump over 'ChatGPT' memoir
The Daily Show
Artificial intelligence (AI) can’t fully understand language and it’s something people are only just getting their heads around.
AI has become ingrained in our daily lives, whether we like it or not. From social media platforms to our phones, AI chatbots are now largely integrated and used daily to simplify tasks and find out information, thanks to tools such as ChatGPT.
But, while humans use language, both spoken and signed, to understand things, it turns out that generative AI can’t, despite how it may seem.
AI pioneer and Nobel laureate, Geoffrey Hinton previously said: “What’s really surprised me is how good neural networks are at understanding natural language – that happened much faster than I thought…. And I’m still amazed that they really do understand what they’re saying.”
But, Veena Dwivedi, a professor in the Department of Psychology and Centre for Neuroscience at Brock University, has argued that this is wrong, first and foremost, because text on a screen is regularly confused with language, but, though they’re related, they are not the same thing.
Dwivedi argues: “Linguistic communication mostly happens face-to-face, in a particular environmental context shared between the speaker and listener, alongside cues such as spoken tone and pitch, eye contact and facial and emotional expressions.”
iStock
To explain further, she posed how the two words, “I’m pregnant”, may be interpreted differently depending on context, for example, between a married couple who have been facing fertility issues, versus a teenage couple experiencing an unplanned pregnancy.
She said: “In each case, the message recipient ascribes a different sort of meaning – and understanding – to the very same sentence.”
Dwivedi also explained that neural networks in AI refer to the algorithms and are not to be confused with biological brain networks in humans.
So, while a computer or an AI chatbot may be able to respond to language in text form, it lacks the emotional context needed for it to truly be able to reach a human level of understanding.
Dwivedi explains that even babies are able to glean context cues in language, and as humans, we use more than just the words spoken to extract meaning and understanding – something AI is unable to achieve.
Why not read…
China's army of robots is worrying experts
Who does AI think is most likely to get arrested? Funniest videos from family to pets
You can now add an AI boyfriend to your photos with this new filter
How to join the indy100's free WhatsApp channel
Sign up to our free indy100 weekly newsletter
Have your say in our news democracy. Click the upvote icon
Top 100
The Conversation (0)