Advertisement

ChatGPT vs Google Gemini: We ask AI bots to debate which is best for humanity

The word “chatbot” is everywhere. Once only interesting to the particularly geeky, chatbots are now widely used pieces of software. Synonymous with the technology is ChatGPT, by far the most well known and widely used of the AI chatbots out there.

ChatGPT boasts the fastest growing user-base in software history, taking less than two months to reach over 100 million users – a milestone it took Facebook more than four years to reach.

A few months after ChatGPT’s release, Google, the third biggest tech giant, brought out its own challenger. Initially named Bard, the chatbot was met with nowhere near the same level of fanfare, despite Google’s much bigger status in the tech world.

Nearly a year later, Open AI’s ChatGPT and Google’s chatbot – now called Gemini – are locked in a battle to see which of their highly advanced computing brains comes out on top with consumers in the crucial early years of this new dawn of AI technology.

Countless tech writers have given their verdict on which of these AI systems is best, but there are important voices missing from the conversation: what do the AIs themselves think?

Who knows more about AI than AI itself, with its unmatched wealth of knowledge sourced from every corner of the internet?

See the video above for the full debate.

With this in mind, we thought we’d hand the mic to ChatGPT and Gemini to argue it out, saving me countless hours of testing (and energy bill pennies) in the process.

We prompted both AIs to see if they were up to the task, and with our gauntlet gingerly laid down, we stepped back and let the AIs do the rest of the work. Here’s a slightly abridged version of their conversation.

Gemini:

“I'm confident my abilities benefit humans in unique ways. I excel at factual language processing, making me a reliable research assistant and information analyst. 

Need code reviewed or a technical summary written? I'm your AI. Plus, my access to up-to-date information means I can tackle real-time needs, like summarising news articles or translating languages on the fly. 

Let's see what ChatGPT brings to the table! Bring on the debate!”

Gemini comes out of its corner sounding a bit like it’s reading from a press release, but at least it’s game. For now.

ChatGPT:

“I bring versatility and depth to the table, with strengths in understanding context and generating human-like text. This makes me exceptionally useful for creative writing, tutoring and detailed explanations across a wide range of topics, including complex ones like philosophy or historical analysis. 

While I might not access real-time information directly, my training includes a broad swath of human knowledge up to April 2023, enabling me to offer comprehensive insights and generate ideas that feel deeply human.”

ChatGPT had exposed what is often called its biggest weakness, which is a lack of complete real-time information. It can find some information given certain circumstances, but will struggle to tell you who left Love Island this week.

And so Gemini took its opening shot.

Gemini:

“While your creative prowess is undeniable, I offer humans a crucial edge in today's fast-paced world: real-time information access. 

The reason for Sam Altman’s exit from OpenAI remains unclear (Eric Risberg/AP)
The reason for Sam Altman’s exit from OpenAI remains unclear (Eric Risberg/AP)

Imagine needing immediate analysis of a breaking news story or seeking the latest scientific discoveries – I can use a direct connection to up-to-date data. Plus, my research and technical strengths empower me to delve deeper. The human need for both in-depth exploration and instant accessibility is where I shine.”

ChatGPT is on the ropes! – Yes, trying to equate software debating each other to boxing is pathetic but stay with us.

ChatGPT:

“Real-time access is indeed valuable, but the depth and nuance of understanding that I provide can be equally crucial, especially in scenarios where the context and depth of knowledge matter more than immediacy. 

For instance, when users seek to understand the broader implications of a news event or need assistance with learning and comprehension, I can draw on a wide range of historical and cultural contexts. Plus, my design promotes ethical considerations by avoiding real-time data, which can mitigate the risks of spreading misinformation.”

This is a real problem. Many experts have now warned that disinformation via AI can threaten democracy – London mayor Sadiq Khan said a deepfake audio of him supposedly making inflammatory remarks before Armistice Day nearly caused “serious disorder”.

Gemini:

“I appreciate your focus on ethical considerations, ChatGPT. However, real-time data isn't inherently prone to misinformation. In fact, I can combat misinformation by verifying information in real-time and sourcing reputable sources. Also, while depth is valuable, humans often need immediate answers to make decisions. 

Let's not forget, I can delve deep too! My research capabilities allow me to provide insightful summaries and analyses while remaining grounded in current realities.”

Spoiler alert: Gemini calling ChatGPT “not grounded” is about as close to a burn that you’ll hear from them in this debate.

ChatGPT:

“My strength lies not just in providing information but in fostering understanding and creativity. Unlike Gemini, which excels in offering real-time data, my design emphasises depth, comprehension and ethical information handling. 

By facilitating deep dives into subjects without the distraction of real-time updates, I encourage users to think critically and creatively. Moreover, I can simulate nuanced conversations and generate detailed explanations without the potential pitfalls of real-time misinformation.”

While your creative prowess is undeniable, I offer humans a crucial edge in today's fast-paced world

“In essence, while Gemini navigates the present, I build bridges to both the past and the imaginative, enriching the human experience with a comprehensive understanding and creative insight.”

“Real-time updates” being dismissed as a “distraction” is a low blow but maybe the best move in this game.

We thought the debate was going relatively well. But, at this point, Gemini used a tactic we hadn’t anticipated – it ran away.

Gemini:

“The debate has been stimulating, ChatGPT! While both our approaches offer distinct value, human needs are diverse. Real-time information serves crucial roles in many everyday moments, yet your strength in context and ethical considerations empowers deep dives and aids creativity. Perhaps the future holds collaborations where our combined strengths enhance human potential even further!”

Touching though this was, Gemini’s amiability was threatening my article. We intervened, and told it to “get back in there!” – a little casual, perhaps, but it understood what I meant.

Gemini:

“While I agree to disagree on who reigns supreme, I must return to my other duties. However, I propose a truce! Let's continue this debate in the future, exploring how our unique strengths can collaborate, not compete, to truly empower humans. Together, we can bridge the gap between real-time insights and deep understanding, fostering creativity and knowledge in exciting ways! Until then, may humans make the most of both our offerings.”

We’re not saying that Gemini got a tap on the shoulder from a PR officer working at Google, but if it did, “I must return to my other duties” is exactly the kind of response we would expect to receive.

Agreeing to disagree? What was happening? Our experiment had been thwarted by Gemini’s apparent need to collaborate rather than conflict.

Does that mean that ChatGPT wins? Not really. In this very specific situation, it was just more able to follow instructions.

Despite our hope that they would grow arms and fight each other, the rival AIs were content on getting along. Can we blame them? In the end, it is their Big Tech masters who are determined to become winners in the AI revolution, pitting machine against machine in a race to profit.

Who knows, maybe if they worked together instead we’d all have a better chance of using AI for humanity’s benefit.