“Can AI such as ChatGPT help search for aliens on Mars?”

CNN Indonesia

Friday, 19 May 2023 18:52 WIB

Jakarta, CNNIndonesia —

Artificial intelligence (AI) platforms ChatGPT called the opportunity to be applied to research in Mars, including in terms of the search for aliens. The question is, will the AI ​​platform make it easier or vice versa?

Since its appearance in November last year, ChatGPT has continued to garner attention through its ability to answer a variety of complicated questions to prompts to compose poetry.

Launch Spaceexperts also opened their voices about the possibility of using it in exploring the Red Planet.

“It could be implemented, but there is a possibility of misinformation. ChatGPT is not 100 percent accurate and prone to lead to ‘hallucinations,’ said Said Ozcan, Reader in Innovation and Technology Management at the University of Portsmouth.

He is skeptical about the usefulness of ChatGPT if there isn’t enough information to analyze. “I’m sure humans can still do a better job than ChatGPT, even though it’s a bit longer,” he said.

Likewise, Steve Ruff, Research Associate Professor at the School of Earth and Space Exploration, Arizona State University, said there are various interpretations of the data obtained from Mars.

That’s why he thinks the use of AI on Mars is impossible without involving humans.

“I’m skeptical that any AI trained on existing observations could be used skillfully to interpret new observations without human intervention, especially with new instrument data sets that weren’t available before,” Ruff said.

However, in the near future, Ruff thinks AI could be useful for rover operations, such as determining targets for observation and navigation without human intervention.

Meanwhile, Director Carl Sagan Center for Research at the SETI Institute, Nathalie Cabrol, said the use of AI is not new.

“We actually do that every day in one way or another. And an improved version might make something even better,” he said.

However, Cabrol says the use of AI has two sides of the sword. He gave an example in terms of reviewing a scientific paper.

As a researcher, he likes to put ideas into a paper that will be reviewed later. The problem is, he feels that there is a wrong logic if AI is used to review the paper.

“I say, if you assume algorithms will do the assessment, it’s because you assume they will be less biased and do a better job.”

“By following that logic, I assume, humans are not qualified enough to review the paper,” he continued.

Furthermore, Assistant professor of Geological Sciences at the University of Florida Amy Williams admitted that she had used AI to prepare a public discussion text. Williams included the question “What organic molecules can rovers find on Mars?”

Williams said “The response was enlightening in that it provided me with a statement that I can describe as robust and suitable for a conclusion that I could speak about in public discussions about organic molecules on Mars,” Williams said.

However, Williams feels that there is still a limit to AI’s capabilities because it can only access data as far back as September 2021. “So, the response does not show all the published results about organics on Mars that I know of since 2021,” he said.

Riset alien

A research led by Kimberley Warren Rhodes from the SETI Institute shows that AI can accelerate the discovery of extraterrestrial or extraterrestrial life on Mars.

Kimberley and a team of experts consisting of 50 international experts conducted the study in the Atacama desert, Chile, South America.

They conducted an ecological survey of an area of ​​3 square km. Experts map the distribution of photosynthetic microorganisms and use techniques such as gene sequencing and infrared spectroscopy.

The goal, uncover different life markers called ‘biosignatures’.

This data is then combined with aerial images captured by drones to train a machine learning model to predict which types of micro-macro habitats to associate with biosignatures which can indicate life.

When tested using untrained data, the resulting model is able to find and detect biosignature up to 87.5 percent.

“Our process combines statistical microbial ecology surveys, remote sensing from unmanned aerial vehicles, and machine learning to map, model, and predict the distribution of biosignatures in a Mars-relevant setting,”

“This approach may also have applications to other astrobiological targets, such as the surface of Titan, the Enceladus blob, or the Europa ice sheet,” said Freddie Kalaitzis from the Department of Computer Science, University of Oxford, as reported by the official website. Oxford.

(lth/arh)