Moderation is the key to lasting enjoyment.
This quote by Hosea Ballou, an American theological writer, fits best with the current AI scenario.
As AI emerges, so do its negative side and people’s concerns.
However, now it seems there is no going back for AI. But balancing its uses with certain responsibilities is becoming the need of the hour.
To deal with this, many AI trends are going on. Various organisations, companies and governments are participating and calling for responsible AI.
This time, YouTube has come forward to deal with its potential threats.
In its blog post titled “Our Approach to Responsible AI Innovation.” YouTube announced that people or content creators would soon have to keep certain rules in place if their content comes through Generative AI.
Although it did not mention the specific date when these rules will actually be rolled out, hints were that they would be there in the coming months of the new year.
Out of those hints, let’s see what changes YouTube is on its way to introducing.
Synthetically created content will be labelled.
YouTube said if you create content through synthetic means, i.e. generating content via AI tools, you will have to disclose this to people.
This is to protect viewers from any misleading information.
Creators may come up with storytelling that depicts some event that never took place or even show any specific third person saying something which, in reality, the person didn’t say at all but is the outcome of the deep fake.
In this scenario, YouTube says it becomes very important to reveal its origins, especially when it is about elections, ongoing conflicts, public health crises, or public officials.
To deal with this, YouTube comes up with three levels of solution
- It plans to add a new label in the description section indicating that content comes via AI tools.
2. If the content is about sensitive topics, then YouTube hints at showing a more prominent label to the video player.
- In case the content is even more sensitive, like showing violence, YouTube will, without any hesitation, end up removing it even if it has a label on it.
Privacy and authenticity will be, above all.
Using the face or voice of an identifiable person in a video without permission will be a privacy issue, especially when it intends to misrepresent their viewpoints. Also, if there is any artificial content mimicking any music, rap, or voice, it puts authenticity at risk.
In that case, the aggrieved personality may request to remove the content. The platform will do so, however, not without evaluating certain factors.
AI for AI detection
YouTube said both people and technology together have already been placed to work as vigilantes, reviewing content on Google from all over the world.
So if there is any violative content, AI classifiers detect it, and human reviewers confirm the same.
But AI now seems to be even more accurate and quick, identifying any form of abuse no matter how new or emerging it is.
With all that, YouTube will penalise abusive content going against policy more rigorously.
Towards even more responsible AI
YouTube is developing various AI tools to help creators. However, they resemble responsible AI- which is what YouTube is striving for.
Both teams and technologies are working to moderate content according to policies.
For this, YouTube said, “We are developing guardrails that will prevent our AI tools from generating the type of content that doesn’t belong on YouTube.”
By chance, bad actors capably bypass these guardrails. YouTube says, “We’ll incorporate user feedback and learning to continuously improve our protections.”
For further protection, YouTube said, “dedicated teams like our intelligence desk are specifically focused on adversarial testing and threat detection to ensure our systems meet new challenges as they emerge.”
For now, this information is what YouTube has revealed. As mentioned, in the coming months, YouTube will place these updates into practice so we will be able to understand them better.
To creators, it may concern creators a little. But from viewers’ and artists’ perspectives, it is all about more transparency, accuracy, and privacy, protecting from misleading or biased information.
We hope you liked the information! To get access to much more interesting news and information, subscribe to our blog.