AI and the issues with deepfakes assessment 2

Using critical digital literacy skills to evaluate online information on issues with AI regarding Deepfakes

I have been applying my critical digital literacy skills when searching for information on issues with AI regarding deepfakes.

What are Deepfakes?  

Deepfakes are videos/audios/images that seem realistic but are actually generated by AI. At first glance they seem realistic, but it is completely fake. There have been cases where these deepfakes have been used spitefully to influences elections and create non-consensual porn. Below is a video by VICE which they posted 10 months ago, it shows a deepfake of Barack Obama, this is a much older deepfake and the ones now are even more realistic then this which is disturbing to imagine. 

SOURCE 

(if source is not accessible: https://www.gao.gov/products/gao-24-107292)

If the video is not accessible click this >>> CLICK ME

(https://www.youtube.com/shorts/Fyzr6jbSI_I)

Why are Deepfakes an issue?

Deepfakes are an issue because it spreads false information, poses a threat to privacy and security while having the power to impersonate anyone. Cybercriminals can create a number of new scams or identity theft with disturbing accuracy with the help of deepfakes. 

An example is a multinational company that got scammed by cybercriminals using deepfakes POSING as their company's chief financial officer. They sent an email for a video call and during "the meeting" convinced to do 15 bank transactions over the course of the week adding up to $40 million dollars.  

SOURCE

 (if source is not accessible: https://www.google.com/url?sa=j&url=https%3A%2F%2Fia.acs.org.au%2Farticle%2F2024%2Fcompany-loses--40m-to-deepfake-video-call.html&uct=1722144955&usg=M_yxyUeFVKkS7bsrwBVjJxxRXdk.&opi=98421741) 

 Here is Addison Rae an American actress, dancer, singer & TikTok influencer who fell victim to deepfakes that created a fake non-consensual porn GIF of her, it has since been taken down. 

IMAGE SOURCE (if link is not accessible: https://powerhouseforensics.com/deepfake-forensics/celebrities-impacted-by-deepfakes/ )

Who is affected by this issue?

Deepfakes can affect politicians , celebrities and big companies. These are the primary victims but  everyone around the world who use the internet/social media can be a victim of deepfakes. Here are some statistics regarding deepfakes. It is said that 60% of consumers have encountered a deepfake video within last the year , 1 in 4 leaders are unfamiliar with deepfakes , 1-10 executives have already faced deepfake threats , 10x increase in deepfakes detected globally across all industries in 2023 and another statistic i found interesting is that only 38% of students have received guidance from their schools on how to identify AI generated images/texts/videos despite a desire for such training expressed by 71% of students. I wanted to include this statistic because it is saddening how 71% have asked for training but haven't been taught, as this would be amazing for their future to identify them in their jobs/everyday life.

(all links for statistics are on the statistic, please click the statistics for each source if not accessible: 

https://eftsure.com/statistics/deepfake-statistics/#section-3)

How does it occur? 

They occur by using a complex process that relies on AI algorithms specifically focused on deep learning. These algorithms analyse thousands of images/videos to learn how to mimic facial expressions, movements, voices etc. 

The chosen critical digital literacy evaluation method is the TRAAP method. I will be using this to evaluate the article I chose. 

Science & Tech Spotlight : Combating Deepfakes << CLICK FOR ARTICLE

 (if not accessible https://www.gao.gov/products/gao-24-107292 )

Timeliness - When was this source published?

Relevance - Does this source meet my information need?

Authority - is the author of this source an expert? how do you know?

Accuracy - is this information accurate? does it match what I found in other sources

Purpose- Why was this written? 

Why did I choose this evaluation method compared to the others? 

I chose this evaluation method because it was most relevant in my opinion. I looked at CRAAP it is much similar , but TRAAP seemed to fit my article better. I also tried out PROMPT but was confused about M which was method. Also the first P which was presentation didn't apply well to my article as its very bland and formal since its a .gov website and there is a lot of words rather then pictures. With all that into consideration I decided to stick with TRAAP as it already fit my article evaluation process.

IMAGE SOURCE: 

(If source doesn't work: https://www.google.com/url?sa=i&url=https%3A%2F%2Fwww.scribd.com%2Fdocument%2F638707545%2Fbrinsfield-slm552-traap-poster&psig=AOvVaw3ge4hddfqOAmNp3HyCKNmU&ust=1722250225608000&source=images&cd=vfe&opi=89978449&ved=0CBEQjRxqFwoTCIiIwZjIyYcDFQAAAAAdAAAAABAE )

Thank you for reading my blog. By Jiselle Tamraz S00368635