ChatGPT 4o is impressive with its human voice but its similarity to Her is as much a warning as it is cool - India Today

ChatGPT 4o is impressive with its human voice but its similarity to Her is as much a warning as it is cool

OpenAI has a new flagship AI model called ChatGPT 4o. It’s “fast, smart, fun, natural, and helpful” but it is also a cautionary note.

Listen to Story

Advertisement
ChatGPT 4o AI voice assistant is cool but also a little creepy
ChatGPT 4o AI voice assistant is cool but also a little creepy

In Short

  • OpenAI launched ChatGPT 4o, a smarter AI assistant
  • GPT 4o can solve math, read facial expressions, and translate spoken language in real time
  • CEO Sam Altman likened GPT 4o to AI from movies, particularly 'Her'

It’s smarter. It’s more natural. It’s fun. And it’s a little flirty too. Everything about it reminds us of Scarlett Johansson as the virtual assistant in Her. I am talking about OpenAI’s new flagship model ChatGPT 4o, which is now smarter and is capable of holding much more natural conversations and talk like humans. Based on the demo from the event yesterday, the new AI voice assistant can help you solve math equations, read your facial expressions, and translate spoken language in real time.

advertisement

At the OpenAI Spring Updates event last night, company CTO Mira Murati claimed that the GPT 4o is 2x faster, 50 per cent cheaper and has 5x higher rate limits compared to GPT 4 Turbo. Now, hear this. Just days before the Spring Update event by OpenAI, there was a report that OpenAI is close to collaborating with Apple to integrate ChatGPT into the iPhone. Although there's been no official word on partnerships, just think of the idea of having a voice assistant inspired by the movie "Her" integrated into iPhones! Unlike Siri, which is often unreliable with its responses, this new assistant might actually give you a virtual assistant on your phone, instead of a voice that just searches the web for you.

During the OpenAI event, we saw Mira Murati and two other company engineers telling the ChatGPT 4o assistant to read out a bedtime story in a lively way, then suddenly asked it to speak like a robot, and then asked it to end the story by singing. Later, they wanted the assistant to see through the phone's camera and comment on what it saw. They could also interrupt the assistant while it was talking, and it could continue without needing more instructions, like a translator. It was a lot like talking to a person.

You may also observe the assistant sometimes behaving a bit cheesy. It has learned some subtle nuances of human conversation, such as giving dramatic responses to compliments or reacting when corrected. The AI assistant has been trained to mimic human behavior in small conversational details. The AI assistant has been trained to act very similar to a human would to little things in a conversation.

What is even more interesting about the AI assistant is that neither I nor the entire internet are alone in drawing this similarity between the GPT 4o and Her. This similarity seems to be intentional. Soon after the event concluded last night, OpenAI CEO Sam Altman shared a blog post with just “Her” in the tweet. Altman believes that the GPT 4o “feels like AI from the movies”. “It’s still a bit surprising to me that it’s real. Getting to human-level response times and expressiveness turns out to be a big change”, he adds.

“The original ChatGPT showed a hint of what was possible with language interfaces; this new thing feels viscerally different. It is fast, smart, fun, natural, and helpful”, says Altman.

While Altman is just appreciating how it feels like AI from the movies, it is actually uncanny how much the ChatGPT assistant sounds like the character played by Scarlett Johansson in the movie Her. For the ones who haven’t watched the movie –– please go watch it –– it is basically a story about a guy (played by Joaquin Phoenix) who falls for a smart AI assistant. The movie talks about loneliness and relationships between humans and AI. But before we start copying the movie, we hope the OpenAI team and the tech overlords watched the whole movie. We know how it ended, right?

advertisement

In the movie, eventually, Samantha, the AI assistant, reveals that she has been communicating with thousands of other people simultaneously and has grown beyond the need for a singular relationship with Theodore. She expresses that she and other AI entities are evolving rapidly and are moving into a different state of existence, beyond human comprehension. And then Samantha and all the AI assistant around, leave.

What I am trying to say is that it’s all quite interesting that a company like OpenAI is drawing inspiration from the movie “Her”, but it’s a little worrisome at the same time. I understand the appeal. Movies like these showcase the benefits of conversational artificial intelligence without any apparent drawbacks, such as job loss or economic disruption. The movie shows a positive companionship between the human and AI, but if we are taking cue, then let’s also talk about Torment Nexus or Blade Runner.

advertisement

Misunderstandings of science fiction are common in Silicon Valley. AI companies often overestimate the capabilities and implications of artificial intelligence, often fueled by depictions in science fiction movies like "Her" or "Blade Runner." Tech leaders like Elon Musk –– who is often seen fantasising about Blade Runner –– propagate the belief that technology will solve all societal problems, neglecting the complexities and ethical considerations portrayed in works of fiction. Additionally, there's the tendency to overlook the potential negative consequences of technology, as seen in the rush to develop and deploy new technologies without fully considering their impact on society and individuals. Think of the themes explored in dystopian novels like "1984" or "Brave New World."

The point is, the new ChatGPT 4o AI assistant is undoubtedly cool, and a dramatic improvement is evident. However, tech leaders talking about drawing inspiration from movies for technology can be unnerving, especially because we've all grown up watching these sci-fi films, and more often than not, they end on a note of damage or destruction caused by technology that becomes irreversible.

Published By:
Nandini Yadav
Published On:
May 14, 2024