Leaked teen social media ban tech trial results don’t support claims

When Australia trial of Teen Social Media Tech trial, the tests claimed that technology could work effectively, the hearing experts wondered why they did not see evidence.

After continuing for a better part of a year, the age assurance technology trial, assigned by the government last month, published 12 results list. There were extensive expressions such as “Age assurance in Australia and may be special, robust and effective ,, but it did not contain evidence that supported them. Supporting data will be included in a report later this year.

Now, a special draft of the report, some experts, the first “pre -findings” exaggerated and exaggerated and exaggerated and exaggerated whether or not the allegations of the hearing has been found in the wrong statement.

Director of the hearing, Age Check Certification Program CEO CEO Tony Allen, the organization will not comment on the results until all results are published, he said.

When asked about various figures in the report – with a few people who see the draft report and based on the discussions that see the sections of the document Crirase – Allen said that he did not recognize the numbers you quote ”and that any report on these numbers will be wrong if these numbers are quoted”.

Trial findings presentation

Last week, 20 experts – stakeholder Advisory Board – a “secret özet draft of the first part of the hearing was given to the 20 experts.

Before that, very little was known about what the test revealed about technology observed to implement the young social media ban.

The decision of the hearing to publish a statement about the findings without including any of the underlying evidence. Crirase Special and publicly open.

Now these experts have seen some of the draft of the report – as “information for potential include the final report ve and its content“ does not represent a final location ” – Crirase He said he had more concerns about relying on the results of the hearing.

It suggests that its broad statements are not supported by evidence, the results are surprisingly based on limited tests, and that the evidence provided may not be as ready as age control technologies.

Draft report

According to several people who see the report, the age estimation technologies tested-the age of someone who is based on the age of someone and other biometric data does not provide much information that predicts that the age-tahmins are closed for an average of 2,69 years.

ABC News reported Some of the technology tested in the experiment estimates the age of people in a 18 -month range in 85% of cases.

The report also said that facial analysis technology performs the worst performance in the first nation populations, which makes it less appropriate in the context of Australia. Worse performance of face recognition and analysis technologies in non -white populations well established.

The draft report says that a person tested the 28 -year -old provision technology provider, a sub -part of the age aid technology, which controls a person’s age against “identity documents or reliable data resources”. (Like a digital fedai that controls your identity before leaving you to a pub.)

Of these, the report could say that the majority was only 7 or above (out of 9) a “level of technology ready” (TRL), at least they had a working show. This means that the majority of the participants have a fully working product ready to be commercially ready to be used, which will be TRL 8 or more.

According to the number of tests listed in the report, each provider was tested 340 times in a controlled school environment. It was tested by the “mysterious shopping”, each of which was a wage hearing participants, each of which was tested outside a controlled test environment, each. (The experts of the hearing were initially told that those who were mysterious shopping were only adults, but since then, the group was told that children also included children.)

Average accuracy for age-birth technologies was 97.05% for mysterious shopping, identity control technologies or other resources either misrepresent the age of one, or about 3% of the time could not confirm approximately 3%.

The report also means that the providers have a “ISO/IEC 27001” certificate for the safety of at least 12 ği, that is, an independent auditor controls their protection for sensitive data.

The instant image of important statistics without trying

Hearing experts are worried about the results

Even in this missing part of the report, some experts question whether the data of the hearing supports comprehensive allegations about the effectiveness of age assurance technologies.

Average error rates, the number of products that are not commercially ready, the number of tests performed, and the limited number of provider with proven data confidentiality management systems, age assurance technologies are ready to use and the claim that it was “special, robust and effective”.

Some criticized the decision of the hearing with the allegations of the preliminary report and the limited information published in this draft, while others pointed out the restrictions of testing an emergency technology in a tight time period decided by the government.

Experts of the hearing are gathered this week to give feedback about the report and the hearing. The latest version of the report will be given to the government by the end of the month.