Pages

Instagram Tests New AI Tools That Are Designed To Verify Users’ Age Through Face Scans

Instagram is entering a new realm of innovation with the announcement of the app conducting tests of unique AI tools that are designed to verify your age.

The company says it is testing out the great benefits associated with AI technology and through that, it hopes to develop an accurate method for age testing using face scans.

Therefore, users can now rely on the computer for age estimation, simply by running their face through a scan.

The fascinating tool has been created by a third-party firm that goes by the name Yoti and so far, things are looking promising.

Obviously, Instagram is well aware of the age restrictions associated with such developments and that’s why the company says you need to be at least 13 for eligibility for an account on the app. Interestingly, we haven’t seen the firm make too much effort towards the enforcement of this policy but we feel now is the time that changes are being made.

Remember, it was until 2019 that we saw the app ask users about what their birthdate was, forget about entering the verification process. But now that we’re seeing stringent measures come into play by different privacy experts, age verification is being dealt with more seriously.

This way, we’re no longer at a loss of being able to differentiate who the adults are from the kids.

As far as the current policy of the app is concerned, you’re asked to verify your age-related details in case you wish to change your birthdate. This can be done by simply adding images of different ID cards. And from now onwards, American users are also going to have a number of options to choose from. This includes vouching through social media and the use of AI for age determination.

For social media vouching, three different followers of a user would be required to confirm the person’s age. Obviously, those being asked to verify need to be above 18 themselves. Also, they will be given a deadline of three days to answer.

The second method is related to the new AI-powered age tool that makes use of machine learning to determine the age.

For those who may not be aware, Yoti is the name given to a leading company that has been used routinely online for age estimation and for verification of identities. Hence, it only made sense as to why they would join hands with a leading social media app like Instagram for this endeavor.

Yoti has also sought approval from leading regulatory bodies in the UK as well as Germany. But how exactly does it work is a question on so many people’s minds.

Well, the company says it uses facial signals to help come to the conclusion of an individual’s target age. Shockingly, the firm says it is still trying to get its fingers on what those signals may be. Perhaps the answer is wrinkles or maybe a look of tiredness or sadness? To be honest, it could be all.

Thankfully, the firm says anyone and everyone can try it out on the firm’s website and it promises not to share any of your details with anyone. You can even see how accurate the results have been in the past so giving it a try wouldn’t hurt.

The results for accuracy determine the error rate for the age calculated in years. And they vary according to your skin color, gender, and your age bracket.

On average, we see that the system may be less exact for the faces of females and those with darker skin color. This includes those below the age of 24 years, where estimates could be a little off by 2 or perhaps 3 years at max.

Surprisingly, one non-profit firm says they carried out the age estimation analysis for its employees in 2020 and found the results to be almost 99% accurate. So, you can definitely count us in on this one because the margin of error is definitely less.

For now, we’re not quite sure how it will work for Instagram and whether or not it’s trustworthy enough as it’s still in the testing phase. But we’ll definitely keep you updated on any more news in this regard.


Read next: Instagram is putting an end to its incentive program on June 30

No comments: