I think that it will not be possible to know if something is AI-generated, because AI is always developing, and will eventually reach near-human intelligence.

• AI is simulation of natural intelligence which has been processed by machines, mainly computer systems. 
• Runs on a series of advanced algorithms and specialized hardware. 
• AI algorithms are instructions that allow machines to analyze data, perform tasks and make decisions. 
• However, what makes it unique yet incredibly dangerous is its adaptable nature, and its ability to take in data and learn from it. 
• There are many issues presented from the idea of AI (privacy concerns, security risks and loss of job opportunities, etc)
• AI has the ability to infer a person’s location and private data
• Could be used in identity theft, which are caused by privacy issues and security risks created by AI.
• AI also leads to job loss, which could lead to high unemployment rates and a lack of human creativity that is needed for the positive growth of the world. 
• As AI develops, it may be impossible to distinguish between AI and natural intelligence. 
• This could prove to be dangerous, with AI perhaps being so close to human intelligence, it can act just like a human being who has extraordinary intelligence. 
• This may lead to a decrease in job opportunities for actual human beings, privacy concerns, security risks and a general a lack of natural human creativity in society.
• This project will experiment with ways to determine what is or isn’t AI created, and how to know what is generated by AI
• AI has the potential to be something revolutionary, yet incredibly dangerous. 
• It can make privacy violations, that it can lead to algorithmic bias and that is can lead to job loss. 
• These days, it is obvious to tell what is AI generated using AI checkers or intuition. But what about in the next 5, 10 or 20 years from today? How advanced will AI be by then? 
• What is the nature of AI? How can we tell if something is AI or not? What can we do to prevent AI from reaching a point where natural intelligence is not needed? Is it possible to know if something is AI generated?
• The thing we don’t know about AI is the most dangerous thing about it: We don’t know how unexplainable and unpredictable it is
• All of these things are crucial in making sure AI can not lead to making a negative impact on society

• My project will experiment with methods to determine between AI and human intelligence.

  It will test the reliability of these methods and prove if it is or isn’t possible to know what is or isn’t created by AI

  In this experiment, I will give a series of human and AI generated pieces of writing or art to a variety of people and AI checkers. I will tell them ways to know between artificial and natural intelligence. Finally, without telling the people or AI checkers which or the pieces of writing or art is created by man or machine, I will ask them what is created by AI or by a human.

  These experiments will lead to the conclusion that it is or isn’t possible to know what is AI generated or man-made.

Manipulated variable: The piece of writing/image presented to the individuals/checker.

Independent variable: The source from which the writing is from (human writer, ChatGPT, Photoshop)

Controlled variable: The group of individuals being asked, AI checker being used

Step 1/MATERIALS: Prepare the following: An article written 100% by AI, An article written 100% by a human, A human-written article with some changes made by AI, An original image, And finally, a different image expanded by AI. You can use ChatGPT, PhotoShop Generative Fill, and original articles and pictures written or taken by real humans. 

Step 2: Ask 3 different people and 3 different AI checkers which one is AI generated and which one is original/human-created

Step 3: Repeat steps 1 and 2 as many times as you feel necessary. The more tests, the better. What to change: The photo/piece of writing presented to the individuals/checkers. What NOT to change: The people being interviewed and the AI checker being used.

Step 4: Record the results. See if the responses from the humans and AI checkers were accurate to if AI created the piece of writing/photo or not. If the results are not accurate, then it proves that we can't know whether something is or isn’t AI-generated. If the results are accurate, then it is safe to say that we can know if something is or isn’t AI-generated.

Total percentage for respondents without hints: 27% accuracy (total: 5/18)

Total percentage for respondents with hints: 33% accuracy (total: 6/18)

Total percentage for AI checker: 41.6% accuracy (total: 10/24)

As shown, neither AI checkers nor human respondents could accurately identify between AI and human text.

After getting results from the experiment and organizing my data, it is clear that human beings struggle to identify the difference between an AI-written article and a human-written article edited by AI.

To keep it concise, the results from my respondents and AI checkers prove my hypothesis correct. It is not possible to identify between humans and AI, well, at least not consistently, as not a single respondent got every single question correct. After getting results from the experiment and organizing my data, it is clear that human beings struggle to identify the difference between an AI-written article and a human-written article edited by AI. This could prove to be bad news for the world since the ripple effects could cause security violations, privacy breaches, and job loss. 

As seen in my experiment, the use of hints did not change the way a human identifies between an AI and normal human text. 

This information is critical in the real world since it showcases that the world has to do more research on AI detection and provide public knowledge on how to identify AI. If this does not happen, AI will develop to a point where it reaches human intelligence. If more research is done, then we may be able to prevent the algorithms of AI to gather enough knowledge to reach a point of human intelligence. 

My hypothesis was correct since it was not possible to consistently tell the difference between AI and human text. Even the AI section software could not consistently determine between each. However, what may have caused this was not using the latest AI model for each article. Some articles were written by the latest AI software, while others were written by the previous model. This may have caused a difference in the results of my experiment. 

https://www.techtarget.com/searchenterpriseai/feature/How-to-detect-AI-generated-content#:~:text=Repetitive%20words%20and%20phrases.,might%20be%20written%20by%20humans. 

https://www.forbes.com/sites/ariannajohnson/2023/06/07/new-tool-can-tell-if-something-is-ai-written-with-99-accuracy/?sh=77755df95ed4 

https://www.simplilearn.com/advantages-and-disadvantages-of-artificial-intelligence-article#:~:text=The%20drawbacks%20of%20AI%20include,human%2Dlike%20creativity%20and%20empathy. 

https://builtin.com/artificial-intelligence/risks-of-artificial-intelligence 

https://decrypt.co/resources/what-are-the-dangers-of-ai 

https://www.scientificamerican.com/article/how-can-we-trust-ai-if-we-dont-know-how-it-works/#:~:text=But%20AI%20systems%20have%20a,expect%20is%20a%20significant%20challenge. 

https://northwestbylines.co.uk/business/technology/potential-dangers-in-artificial-intelligence-can-we-control-it/

(https://www.semrush.com/blog/how-to-detect-ai-written-content-and-plagiarism/)

(https://originality.ai/blog/how-to-detect-ai-generated-articles)

https://www.history.com/topics/inventions/albert-einstein 

https://www.nobelprize.org/prizes/physics/1921/einstein/biographical/ 

https://brainly.in/question/57140224#:~:text=Throughout%20the%20book%2C%20Harry%20learns,his%20next%20year%20at%20Hogwarts. 

https://www.sparknotes.com/lit/harrypotter/ 

https://hosted.lib.uiowa.edu/histmed/plague/ 

https://my.clevelandclinic.org/health/diseases/21590-bubonic-plague

https://en.wikpedia.org/wiki/Bird_of_prey 

https://www.123helpme.com/essay/Raptors-Birds-of-Prey-203919

https://contentdetector.ai/

https://gptzero.me/ 

https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence 

https://www.tableau.com/data-insights/ai/algorithms

https://blog.hubspot.com/marketing/ai-algorithms#:~:text=AI%20algorithms%20are%20instructions%20that,performs%20work%20on%20specific%20algorithms.

https://www.reuters.com/legal/legalindustry/privacy-paradox-with-ai-2023-10-31/#:~:text=The%20technology's%20potential%20to%20infer,that%20demand%20immediate%20proactive%20solutions. 

• I would like to thank my mother for helping me a lot in assembling the trifold
• I would also like to thank my whole family for taking a bit of time off of their day to respond to my survey