• Reference: 2017/3629
• Question by: Siân Berry
• Meeting date: 14 September 2017

What is the retention period for footage and images recorded by the cameras involved in the automatic facial recognition capability at Notting Hill Carnival? Please specify if there are different retention periods for different sources of images. What are the criteria for further retention of images beyond this period?

• Reference: 2017/3628
• Question by: Siân Berry
• Meeting date: 14 September 2017

This year (2017) was the second time that the Metropolitan Police Service used automated facial recognition at Notting Hill Carnival. What consultation was carried out before the use of this intrusive tactic? What are your measurements of success for this capability?

• Reference: 2017/3627
• Question by: Siân Berry
• Meeting date: 14 September 2017

What criteria were used to place people on the 'watch lists' for a) human recognisers, and b) automatic facial recognition software, to monitor the attendees of Notting Hill Carnival 2017? What was the total number of people on a) both, and b) each of these watch lists?

• Reference: 2017/3626
• Question by: Siân Berry
• Meeting date: 14 September 2017

How many false positive matches highlighted on the automatic facial recognition software were identified after being checked by police officers a) checking against live images, and b) verification after further action was taken?

• Reference: 2017/3625
• Question by: Siân Berry
• Meeting date: 14 September 2017

How many false positive matches were made by the automatic facial recognition software used at Notting Hill Carnival in 2016 and 2017? Please provide a breakdown of the different age groups, gender and ethnicity of these false positives (with ethnicity also broken down by gender).

• Reference: 2017/3624
• Question by: Siân Berry
• Meeting date: 14 September 2017

How many of the positive matches identified using automatic facial recognition resulted in further action by the police? Please provide a breakdown of the total number of individuals identified who were a) stopped by police, b) stopped and searched, c) detained, d) arrested, e) cautioned, and f) other.

• Reference: 2017/3623
• Question by: Siân Berry
• Meeting date: 14 September 2017

How many positive matches were identified using automatic facial recognition at Notting Hill Carnival in 2017? Were all of these positive matches checked by a police officer before being considered for further action?

• Reference: 2017/3622
• Question by: Siân Berry
• Meeting date: 14 September 2017

What was the total automatic facial recognition operational capacity used a) inside, and b) outside the Notting Hill Carnival footprint in 2017?

• Reference: 2017/3621
• Question by: Siân Berry
• Meeting date: 14 September 2017

Which of the front counters and Metropolitan police stations on the list included in MOPAC's Public Access and Engagement Strategy consultation document are CitySafe Havens and what is the current total number for CitySafe Havens in London?

• Reference: 2017/3620
• Question by: Siân Berry
• Meeting date: 14 September 2017

MOPAC's Public Access and Engagement Strategy consultation document, published in July 2017, included a table showing Metropolitan Police Service "Front counters set to close and remain, with average daily crime reports" recorded during May 2017. Could you extend this table to list the number of other reasons members of the public used these front counters, including reporting missing persons, leaving or collecting lost property, presenting documents and any other recorded reasons to visit the police that are not reporting crimes?