• topic:Dealing with privacy infringements as we use AI to prevent crimes

• primary questions:
Situations where AI-aided crime detection/prevention has been put into use or will be put into use?
Specifically, how do risks of privacy leakage occur as AI try to prevent research?
Cases of privacy infringement?
Are there any ways to prevent such leakage? And their pros & cons?
Current laws regarding privacy protection?

[Article]Artificial intelligence and crime: A primer for criminologists

• Three kinds of AI crimes:

AI can serve as a potent tool for ‘malicious’ criminal use by expanding and changing the inherent nature of existing threats, or by introducing new threats altogether.

These same adversarial techniques are also being used in the context of activism and resistance against pervasive surveillance culture(s). In Belgium, researchers designed an adversarial image which, if printed and carried around, rendered a person invisible to AI computer-vision systems.

• Policing uses of AI

First, AI promises (or threatens) the expansion of highly granular digital photography, described succinctly by the ACLU as the ‘dawn of robot surveillance’.

Developments in data storage, along with advances in AI-enabled automatic video analytics, can turn passive, scatter shot monitoring into an ever-more granular, comprehensive, and searchable surveillance record.

Some have argued that facial recognition is categorically different from other forms of surveillance and should be subject to an outright ban (Hartzog and Selinger, 2018). The ACLU, likewise, has expressed concerns that such algorithmic systems are untested, discriminatory, and subject to abuse.

a single drone overflight of a protest could in principle enable authorities to compile a list of all attendees.

GPS chips were embedded in the golf carts, and the flowerbed areas geo-fenced to ensure that golf carts would shut down when approaching the flowerbeds. This example of a fundamental shift to a non-normative regulatory modality for controlling human behaviour, nicely illustrates how, in the future, AI technologies will come to facilitate the sublimation of policing architectures.

[Article]AI Use in Criminal Matters as Permitted under EU Law and as Needed to Safeguard the Essence of Fundamental Rights

• AI is a necessary instrument as information mounts and the world’s complexity increases

• benefits:
• break down obsolete traditional instruments
• make justice a inter-disciplinary system
• management of information
• (author believe) help with the protection of fundamental rights

• Currently, protection of fundamental rights is far from ideal. Thus we need to focus on legislation so that AI would not harm these rights and could hopefully protect them

• AI-evidence

It is expected that as AI shall become a constant component of society it will start to produce content that shall represent evidence.

AI-driven evidence can be a neutral read of reality(coding can be improved) [REALLY?]

• AI-decisions

Formally, the legitimacy of the criminal investigation derives from the accountability of those that investigate.

AI-provided arguments could influence the judge of human

• Legislations
Not very relevant.

[Article]Algorithmic State Violence: Automated Surveillance and Palestinian Dispossession in Hebron’s Old City

Under surveillance capitalism, protected consumers of digital technologies are coerced into producing data and feeding novel configurations of corporate power.

Samira explained how the new surveillance system was encroaching on this last domain of autonomy.

Soldiers regularly entered without warning to maintain the surveillance system.

They act as if we’re not even there.

the new technologies rationalized rather than reduced the army’s intrusion on her life.

I just don’t feel secure inside my own home. And this was the last place I felt secure.

Blue Wolf was not just stealing civilians’ personal data, but also the viability of a dignified life itself if they remained in their homes

These dehumanizing forms of policing and self-policing strengthened the Israeli military’s hold on Palestinian life while offering a valuable resource for a burgeoning entrepreneurial class of army trained developers and corporate CEOs.

[Article]Measuring the Acceptability of Facial Recognition Enabled Work Surveillance Cameras in the Public and Private Sector

• transparency (36% of comments)
regulations on the use of the data collected, knowledge of locations of camaras
💡
keep a record of the usage of video clips

• administrative and safety concerns (34% of comments)
theft prevention is thought of as a legitimate use of such camaras. No performance monitoring

• surveillance and authoritarian concerns (20% of comments)
suspicious of employer’s intention, worries about privacy infringements
“Big Brother watching you all the time”

• some are not against electronic monitoring of performance (6%)

[Article]Privacy and data protection in the surveillance society: The case of the Prüm system

• Two-step approach to exchange DNA information

In step 1, Member States confirm whether or not the DNA profile of their database matches the profile that exists in another Member State's database. After a hit/match confirmation, this will trigger step 2 if requested, i.e. the exchange of further information about the DNA's owner.

The information exchange is performed in accordance with the data protection law of both countries as well as the minimum criteria established by the European Commission for data protection

[Webpage]Artificial Intelligence Regulation Threatens Free Expression

a risk-based approach acknowledges that most AI applications, especially those involving speech and expression, should be considered innocent until proven guilty

The market for AI tools, however, has an important distinction from social media that can encourage greater expression—the absence of network effects.