With AI technology breaking barriers at a speed that most of us are still fascinated by, a new survey is highlighting how brilliant the technology is at confusing people.
ToolTester’s new report speaks about how writers should be on alert because tools like ChatGPT and others in the field of AI are producing content as well as normal humans. And it’s being done at a level where the average American adult has trouble distinguishing AI content from others.
We’re sure that’s not the type of feedback any professional writer would want to hear, considering the great amount of effort they put forward in perfecting their craft.
ToolTester recently carried out two different surveys that took into consideration text generated by humans and those produced by AI tools and edited by humans. And not to many people’s surprise, it was the ChatGPT technology’s 3.5 that was used to get the task done.
After that, another survey was done by taking the latest generation of ChatGPT into consideration. Again, a similar comparison was made, making use of similar prompts. And that’s when we saw more than 50% of the respondents thinking ChatGPT’s 3.5 copy had been generated by the likes of a human. This number was 63.5% and it made use of GPT-4. These results proved how GPT-4 technology at much more likely to convince or trick people than 3.5.
The study also proved how copies generated from the likes of AI technology are so much more difficult to detect, depending on the format of writing involved. A lot of the content related to health did the best to trick people across the board and it appeared as if humans had written them.
This just goes to show how so many people believed that text from AI was actually written by humans, even though that was not the case. Meanwhile, those people who are more familiar with the use of AI tools are definitely less likely to turn into a fool. They end up being misled on most occasions. And those that never use such generative AI tools found it even harder to see the difference between AI content and that produced by humans.
With age, you attain wisdom and that again is worth a notice. People who were above the age of 65 are the ones most likely to identify AI content from human writing. Meanwhile, those between 18 and 24 are likely to feel AI writing is produced by humans.
It does come down to trusting more content publishers against others too. Today, a lot of people have faith that leading media content outlets are going to have articles produced by humans and so many aren’t keen on abusing that. So yes, trust does go a long way here as well.
Read next: Stanford Just Released a Report on the State of AI, Here’s What You Need to Know
by Dr. Hura Anwar via Digital Information World
No comments:
Post a Comment