Summary: The new research shows that AI can create internet memes. The investigators compare the man-made memes, groups of AI, and AIs, using ancient templates such as DOGA and Futurama Fry.
While the memes produced by AI earned the highest score on average, high mememes from the people, in collaboration – AIs very good in formation. The findings suggest AI can help with content content, but one's understanding is still important to break the jokes.
Key facts:
- Ai Memes is a scale on average: The fully produced memes are deteriorating in the average estimates.
- The people are on top: The most fun-made meme made by people, not the equipment.
- The best worldwide: Human-AI cooperation was very old and widely shared.
Source: Rifle
Can AI make fun of? The new research shows the artificial intelligence that can create funny mobile memes as those man-made. But when it comes to gags they really link to viewers, people are still in the edge.
Investigators from Kth Royal Institute of Technology, and TE Damamstadt has just conducted the first high quality lesson to check that people and AI work together to create online memes.
The team compares three groups: People work alone, people create a cut-off language model (llm), and the memes that produces the memes completely.
Participants make up the memes using ancient templates such as Dog, Futurama Fry, with Boromir's icon “person simply …” line. The second group of about 100 people have been measured by the memems wisely, humor, and reconstruction.
The investigators discovered that on average, invertis were fully organized by AI higher than by a person or group groups. But the most effective memes are telling different stories: People were very happy, and Human-AI cooperatives came out.
Published to ACM Digital LibraryThe paper was presented at a 2025 nations' conference with user wise in Cagliari, Italy.
“AI is right in creating more ideas,” said the Student writer, Zhikun Wu, in the king's elections in the KT Royal Institute of Technology. “But the size means the quality.”
You can draw from large dataset, AI models can produce content that attracts many listeners, the authors write. But many limited memes are created by human involvement, suggesting AI models primarily that produce primarily “solid but actions.”
“The good results came when people rejected and shining what Ai produced,” says it.
AI participants who work with AI generally produce other ideas and report a minor effort – but many are fully involved with the program. Few other than a half linked to AI many times, and only hand use it in Iteratively. This limited use may have created the ability to do true.
Studies highlight an important challenge for Human-AI: While AI can produce the content that attracts broader audience, the intelligence is still important to contained deeply content – especially jokes, said.
“Just four jokes about robbery,” says it. “It's about amazing, the essence of cultures, and the emotional nuances Ai are not fully understandable.”
Investigators say that future tools should be better supported, creating dialog-based fees, to help users stay connected to their work while increasing their views. In other words, the systems should not only produce content, but help people to form a purpose.
“While AI can increase productivity and produce the content that attracts many audience, human intelligence is in trouble to create the content that connect the most in some areas,” writes the authors.
In this regard the Instructions Research for intelligence
The author: David Callahan
Source: Rifle
Contact: David Callahan – Kth
Image: This picture is placed in neuroscience matters
Real Survey: Closed access.
“Man is not only a meme: To check the unity of the unity between the people and the people in the generation of fun” Bhikun Wu et al. an arxiv
Abstract
Man is not only a meme: To explore the unity of the unity between llms and the people in the generation generation
The partnership is shown to enhance your intelligence, leading to new and effective results.
While the previous research skills for major languages (llMs) to act as creative partners such as typing poems or create rich llMs and customs remain an open question.
To look at the gap, we conducted the user's ability to test the power of llms in the combination of memes – a funny manner and special cultural traditions of creation.
We conducted the user's study with the three teams of each participant: Only a member of the meskin.
We examined the quality of the memems produced using the crowds, with each tactile meme, humor, and formation. Our results have shown that the LLM help has increased the number of ideas made and reduced the effort they are engaging.
However, it did not improve the quality of the memes when people worked with the llm. It is interesting that the fully created memes is ai made better than the MESES workers with all people and Human-Ai in all places.
However, when you look for the most active memems, the people did the people were better jokes, and a person's cooperation – AI rising from scratches and early agencies.
This findings highlight the difficulty of Human-AI in creative activities.
While AI can increase the product and create content that attracts comprehensive audiences, one's intelligence is always important to the content content at a deep level.
