Since the creation of automated machinery, man versus machine has been an emotive conversation topic. The latest iteration in this ongoing conversation involves the development of ChatGPT and other generative AI.
These intelligent models are now advancing into areas previously seen as solely the preserve of humans – such as art and writing. While organizations like the Associated Press have been using artificial intelligence to create sports reports and other written content for many years, ChatGPT’s generative capabilities are available to all, revolutionizing how people work with words.
Its rise has even led some to speculate that AI will replace human writers. But how do the two really compare? To find out, Relative Insight went a bit meta. We asked ChatGPT to generate a range of content on the topic of artificial intelligence. This encompassed everything from what it is, its different elements, the ethical issues which arise from it and its current and potential uses across a range of sectors.
Then Relative Insight gathered content written by humans covering the same areas of AI. To minimize the possibility that ChatGPT would draw information from the selected human sources, we only used content from 2022 and beyond – after the September 2021 cutoff date which ChatGPT was trained on.
Using the Relative Insight text analysis platform to discern and quantify the differences between man and the machine, the latest battle of AI versus human contained both expected and surprising insights.
Who’s using AI: Businesses or researchers?
One difference between AI versus human content was around who’d use applications of artificial intelligence.
ChatGPT focused on what researchers and programmers would do with the technology. It referenced ‘researchers’ 6.3x more and ‘practitioners’ infinitely more – meaning humans didn’t use the word at all. It was also 1.7x more likely to reference words relating to ‘programming’, such as ‘system’, ‘data’ and ‘algorithm’. This highlights that ChatGPT’s focus was technical, rather than adaptable.
“As computer vision continues to advance, it will be important for researchers and practitioners to address these issues and ensure that the technology is used in a responsible and beneficial way.“
Humans associated AI usage with businesses. They were 8.1x more likely to use words related to the topic, including ‘companies’, ‘entrepreneurship’ and ‘operations. Humans also talked about the use of AI ‘tools’ 9.9x more and referred to the ‘value’ these brought to organizations infinitely more.
“What entrepreneurs and managers should be thinking of – come up with business models that build on humans and enhance the value of human labour and keep humans in the loop, or better yet, humans in charge.“
It’s clear that, rather than researchers helping to chart the course of artificial intelligence, humans expect business needs and applications to drive the future of artificial intelligence.
AI sees the future as evolution 9.9x more
While both humans and ChatGPT discussed the future of artificial intelligence, there was a clear difference in the language they used. Humans were 6.7x more likely to nebulously talk about the ‘future’. They also used the phrase ‘going to’ infinitely more, however, this was to speculate what was next in the world of artificial intelligence.
The language humans use indicates they view AI through the lens of a start and end point, rather than a journey.
“Ultimately it’s going to be beneficial not just for the companies and for society, but for the individuals themselves.“
In contrast, ChatGPT was more likely to describe AI in evolutionary terms. It used words related to this 9.9x more overall, such as the words ‘evolved’ (19.3x), ‘developing’ (11.6x) and ‘continues’ (4.6x). It also highlighted key landmarks we can expect on AI’s evolutionary journey, with the phrase ‘potential benefits’ overindexing infinitely more in ChatGPT’s content.
“As the technology continues to evolve, it’s likely that we will see even more innovative applications of AI in sports in the years to come.“
It’s intriguing that this element of AI versus human content sees ChatGPT examine AI through a biological lens, while humans are the ones focused on the end result in a logical, machine-like manner.
Are humans in danger or are they the danger?
Both sets of content took an in-depth look at the possible drawbacks of artificial intelligence. However, when Relative Insight identified and quantified the linguistic features unique to each, it’s clear that humans may continue to be the biggest dangers to themselves.
ChatGPT was open about the ethical issues raised by artificial intelligence. It used words related to ethics 2.3x more, including the word ‘ethical’ (4.3x), while also addressing ‘concerns’ (9.3x) around AI. These included ‘social concerns’ and ‘job displacement’ – with both phrases appearing infinitely more in ChatGPT-generated content.
“Machine learning has the potential to transform many industries and enable new innovations, but also raises ethical and social concerns around issues such as bias, privacy, and job displacement.“
Humans were more likely to draw dividing lines between themselves and machines. They used personal pronouns 6.7x more overall, including ‘them’ 13.8x more, ‘we’ 2.8x more and ‘us’ infinitely more.
“Whether we embrace these emerging technologies or resist them, they’re undoubtedly here to stay and will continue to play an increasingly important role in all of our lives.“
This focus on division could set the tone as to how humans view even more ‘lifelike’ AI as it continues to evolve.
Quantifying differences between AI versus human content
Relative Insight’s text analytics tool successfully identified key variations between content generated by ChatGPT and written by humans. In this example, the words related to the subject of artificial intelligence. However, you’re able to quantify the differences between human and AI-generated text across any topic.
As more brands and agencies turn to ChatGPT for text generation, there will be greater convergence in what they produce. Leveraging the efficiency of AI-generated content while still maintaining a unique brand identity will be a challenge all brands have to tackle. Relative Insight’s text analytics tool is uniquely placed to support you with this dillema.