The power of disclosure, consent and platforms has become a newly revitalized battlefield with the rise of AI.
The issue has recently come to the forefront as YouTube makes a controversial decision to use “improvements to Unblur, Denoise and Clarity” of some of the content uploaded to the platform using AI-powered tools. This was done with the consent or even knowledge of the relevant content creators. Viewers of the material knew nothing about YouTube intervention.
Without transparency, users are limited in their request to identify AI-edited content, not only to identify it. At the same time, such distortions have a history well before today's AI tools.
A new kind of invisible edit
Platforms such as YouTube are not engaged in subtle image manipulation first.
For decades, lifestyle magazines have been taking “airbrush” photos to soften or sharpen certain features. Not only is the reader not informed of the change, but in many cases the celebrities in question are not. In 2003, actor Kate Winslet angeredly accused the British GQ of change her cover shot.
Also, more and more public are showing their desire to edit images before posting on social media. This makes sense. A 2021 survey of photos posted by 7.6 million users on Flickr found that filtered photos are more likely to gain views and engagement.
However, YouTube's recent decisions indicate that users may not be in the driver's seat.
Tiktok faced a similar scandal in 2021. Some Android users have noticed that “beauty filters” are automatically applied to posts without consent or disclosure.
This is especially concerning as recent research has found a link between the use of ticutok filters that improve appearance and self-image concerns.
Private changes will also be offline. In 2018, the new iPhone model was found to “smooth” the skins of users using a feature called Smart HDR (high dynamic range). This was later described as a “bug” by Apple and reversed.
These issues also clashed in Australia's political sphere last year. Nine News has released an AI-corrected photo of Victorian MP Georgie Purcell, who exposed her midbow, while covered in the original photo. They did not tell viewers that the images they used had been edited with AI.
This issue is not limited to visual content. In 2023, author Jane Friedman discovered that Amazon was selling five AI-generated books under her name. Not only did they not her work, they also pose a risk of serious reputational harm.
In each of these cases, changes to the algorithm were presented without disclosing them to those who saw them.
Disappeared disclosure
Disclosure is one of the simplest tools that must adapt to reality through increasingly changing AI.
Research shows that companies that are transparent about using AI algorithms are more likely to have a significant role for users and AI systems with early trust in the company.
Users have demonstrated a decline in trust in AI systems globally, but also show increased trust in AI using themselves, including the belief that it will inevitably improve.
So why do companies use AI without disclosing it? Perhaps because disclosure of AI usage can be a problem. Research has shown that AI use consistently reduces trust in the person or organization involved, but it is not as if it was discovered that AI was used without disclosure.
Beyond trust, the impact of disclosure is complex. Research has shown that disclosures of false information generated by AI are unlikely to make the information less persuasive to the audience. However, people can be hesitant to share content, fearing spreading misinformation.
A voyage to the unknown generated by AI
Over time, it becomes difficult to identify Commentected and manipulated AI images. Even sophisticated AI detectors are a step behind.
Another major challenge in the fight against misinformation – the problems exacerbated by the rise of AI – is confirmation bias. This refers to a tendency to be less critical of media (AI or others) that confirm what users already believe.
Luckily, we have resources at our disposal. Young media consumers in particular have developed strategies that can counter the tide of misinformation on the Internet. One of these is a simple triangulation. This includes looking for multiple trusted sources to check the news.
Users can also curate their social media feeds by intentionally preferring or following people and groups they trust, while excluding lower quality sources. However, platforms like Tiktok and YouTube leaning towards endless scrolling models that promote passive consumption rather than tailored engagement, so they may face a difficult battle.
The decision to change the creator's videos without YouTube consent or disclosure is likely within the legal rights of the platform, but is likely to put users and contributors in a difficult position.
And given previous cases from other major platforms and what characterised power digital platforms enjoy, this is probably not the last one.
Timothy Koski receives funding from the Australian Research Council.
/Commentary of the conversation. This material of the Organization of Origin/Author is a point-in-time nature and may be edited for clarity, style and length. Mirage.news does not take any institutional position or aspect, and all views, positions and conclusions expressed here are the authors alone.
