The Horrors of Face Sharpening Tools in Documentaries

The Rise of AI in Image Editing

Artificial Intelligence (AI) has revolutionized many sectors, including image editing. Tools like Photoshop and Facetune have integrated AI to enhance photo editing capabilities. However, the use of AI in image editing has not been without controversy. Recently, a tweet highlighted the negative impact of a face sharpening AI tool being used excessively, especially in documentaries. This tool, which is intended to enhance facial features, often results in unnatural and disturbing images.

The Ethical Implications of AI in Media

The ethical implications of AI in media are vast and concerning. The use of AI-generated imagery, such as deepfakes, has raised significant concerns about misinformation and manipulation. According to a study by the Center for Countering Digital Hate, deepfake images about elections are rising by 130% per month on average on X (formerly Twitter). This surge in AI-generated content has the potential to mislead voters, undermine trust in information, and increase political polarization. For more details, you can refer to the Fake images made to show Trump with Black supporters highlight concerns around AI and elections.

The Impact on Public Trust

The widespread use of AI in image editing and generation has led to an erosion of trust in visual media. The controversy surrounding Kate Middleton’s photo editing is a prime example of this. The edited photo sparked significant discussion and debate on social media platforms, leading to the Associated Press issuing a ‘kill order’ to remove the picture. This incident highlights the potential for AI-generated content to spread misinformation and deepfakes, further eroding public trust in visual media. For more insights, you can read about Kate Middleton’s photo editing controversy.

The Accuracy of AI Tagging

Meta’s recent attempt to label real photos as ‘Made with AI’ has also faced criticism. Photographers have reported false positives, with real photos being tagged as AI-generated. This issue reflects the growing concern and need for reliable AI in content moderation and identifying AI-generated images. Pete Souza, a former White House photographer, suggests that basic photo editing might be triggering the AI detection. This highlights the issues with the accuracy of Meta’s AI tagging. For more information, you can refer to the TechCrunch Minute: Meta is labeling real photos as ‘Made with AI’.

The Threat of AI-Generated Misinformation

AI tools from companies like OpenAI and Microsoft have been found to generate misleading election images. These AI-generated images have the potential to spread disinformation and undermine election integrity. This is a significant concern for the general public, policymakers, and technology companies. For more details, you can read the article on OpenAI, Microsoft AI tools generate misleading election images.

The Limitations of Visual AI Models

Visual AI models, despite their advanced capabilities, have limitations. Research has shown that these models often lack actual visual understanding of what they are ‘seeing’. For instance, they can correctly identify a five-circle image but fail miserably when it comes to five pentagons. This is because these models have a five-circle image prominently featured in their training data, such as the Olympic Rings. This lack of visual judgment raises questions about the reliability of these models in real-world applications. For more insights, you can refer to the article on Visual AI models might not see anything at all.

AI-Powered Scams

The rapid advancement and accessibility of AI tools for media generation have also led to the rise of AI-powered scams. These scams include voice cloning for impersonation, personalized spam emails, identity theft, and blackmail with AI-generated imagery. The possibility of blackmail using AI-generated intimate images is particularly terrifying. These image models can produce realistic bodies and attach them to any face, making it easy to fool victims. For more information on how to protect yourself from such scams, you can read the article on AI-powered scams and what you can do about them.