Do you know if what you are watching, hearing, or reading is real? In today’s era of AI generated content and deepfakes can you always trust what you see, hear, or read? Increasingly AI and algorithms are utilized to create and flood social media with content, images, and recordings geared towards stirring emotions and support for certain political and social beliefs and causes. Deepfakes were used in the last two presidential election cycles. They are the tool of scammers to fool you into contributing to a fake cause or to get you to share your or your company’s personal information.
So, what is a deepfake? A deepfake is a type of synthetic media created using AI technology which can generate or alter images, videos or audio to depict real events. They allow a seamless integration of a person’s likeness or voice into content they did not consent to or participate in creating. And those AI generated images are getting better every day. Don’t believe me? Meet Tilly Norwood, the first totally AI actress.
How about AI generated content? I’m sure a lot of us have used and created our own AI generated content, if we’ve ever used ChatGPT, Microsoft’s Co-Pilot or Google’s Gemini to help us polish a report, letter or email, or to create an image to go along with reports or articles. AI generated content is digital media like text, images and audio produced by AI models using advanced algorithms and machine learning to create new content in response to prompts or queries to mimic the human mind.
The problem is when these deepfakes and AI generated content’s goal is to create anger and divide people or to spread misinformation or perpetrate fraud. We need to be on the lookout for:
- Media content that tries to manipulate us.
- Content that seeks to spread misinformation to erode trust and spread fake news.
- Content that is meant to impersonate an actual person or entity for the purpose ofidentity theft
- ruining that person’s reputation, oras a phishing scheme to gain private information.
So how can you spot deepfakes or AI generated or altered content to avoid being fooled? There are AI detection tools that can be used or you can look at the metadata to show when and where the image or video was posted. AI manipulated images often lack that metadata content or show signs the original post was edited.
For images and videos:
1.) Check for unnatural movement, such as unnatural blinking or blank stares. Or robotic jaw or neck movements.
2.) Do they show a lack of emotion?
3.) Are the lighting and shadows inconsistent?
4.) Are the words spoken out of sync with their mouth or lip movement?
5.) Is their skin too perfect or shiny and plastic appearing?
A reverse image search may help you to find where the image used in the post was taken from. Does the content in the new post match the date, time and subject in the original post?
For audio:
- Listen to the voice. Is it monotone or flat?
- Are the words choppy?
- Does the recording have abrupt pauses or unnatural repetitions in the speech?
- For any social media posts or content, visual, audio or written, you need to check the source.
- Is it a trusted source or is the source a known satire site?
- Does it seem to be biased or unrealistic? If it seems questionable you need to question the motives of the post and the person who created the post.
- Investigate the credibility of the website and the author and their affiliations. It can be as simple as typing into your browser (or even you AI site) is ABC Merchandise legit?
- Cross reference the content with multiple sources to confirm it is real and from a reputable source.
- Is the information presented in the content under question consistent with the information presented by other sources
Staying ahead of AI generated and misleading posts will become harder as Tilly Norwood has shown us. But if we want to protect ourselves, our families, and our companies we will need to be on the lookout for AI manipulated content and take steps to be sure that what we’re seeing, reading or hearing should be believed.
About the Author
Brenda McDermott is a Member-at-Large from Region V. She has been a member of IAIP since January 2000. She is a past International Rookie of the Year, Past International Claims Professional of the Year, Past Risk Management Professional of the Year and the 2013 International Confidence While Communicating Winner. She is a past local President, Council Director, Past Region V RVP, Past Region V MAL Director and Marketing Director. She has been a community cafe presenter and is a certified trainer for IAIP. She is currently serving as the chair of the Marketing and Publications Task Force for the 3rd year. She is a Workers’ Compensation Major Case Claims Specialist handling the most catastrophic and complex claims in multiple jurisdictions for The Hartford.