Stop in the Name of AI!

Author(s): Alex Moltzau

Originally published on Towards AI.

A Quick Look into the 2019 AI UN report U+007C Towards AI

A Quick Look at the UN Report on Artificial Intelligence and Robotics for Law Enforcement

Whenever I mention ethics in relation to artificial intelligence (AI) the case of using algorithms in US law enforcement persistently appears as an example. As such, I thought an article on this topic was long overdue. Perhaps not focusing on the specific case popularised and often discussed. Instead, I decided to start with a quick look at a UN report on the topic and summarise the key findings.

Background for the report

In March this year (2019), a new report called Artificial Intelligence and Robotics for Law Enforcement was released by the United Nations Interregional Crime and Justice Research Institute’s (UNICRI), Centre for Artificial Intelligence and Robotics, and Innovation Centre of the International Criminal Police Organization (INTERPOL).

This report is based on insight from the first UNICRI Global Meeting on the Opportunities and Risks of Artificial Intelligence and Robotics for Law Enforcement held in Singapore on 11–12 July 2018. Much has happened since then. In fact, last week the second INTERPOL — UNICRI Global Meeting on Artificial Intelligence for Law Enforcement was held on 2nd–4th of July 2019 as part of INTERPOL World.

INTERPOL World is a global co-creation opportunity which engages the public and private sectors in dialogue and fosters collaboration to counter future security and policing challenges.

Insight from the report

I have decided to make a small selection of bullet points from this article based on what I found interesting, however, you may naturally select other aspects or felt that I missed something crucial. If that is the case you can always tell me in the comment section and give me your opinion.

• Criminals are exploiting technology and INTERPOL-UNICRI must respond. Expertise gaps must be filled.
  It is already well established that criminal groups are not reticent about exploiting technology, having, for instance, capitalised on mobile phones and global positioning system (GPS) devices from their earliest days and, more recently, turning to the dark web and cryptocurrencies and exploiting cyber vulnerabilities. (p.iii) It is also important to advance understanding of and prepare for the risk of malicious use of AI by criminal and terrorist groups, including new digital, physical, and political attacks. Possible malicious uses include AI-powered cyber-attacks, the proliferation of fake news, as well as face-swapping and spoofing tools that manipulate video and endanger trust in political figures or call into question the validity of the evidence presented in court. (p.vi)
• It is worth keeping an eye out for Anita Hazenberg the Director, INTERPOL’s Innovation Centre and Irakli Beridze Head, UNICRI Centre for AI and Robotics. It will be fascinating to see the topics they discuss and the research they release going forward, surely two important voices in the discussion on AI and security.
• Certain use cases are being implemented yet there are many more in other stages of development. The INTERPOL Police Technology and Innovation Radar are keeping an eye out for new use of technology (mostly by criminals).
• The safety community present at these events seems interested in engaging in discussion on the ethics of AI. Part of the challenge with deciphering the ethical use of AI and robotics is that law enforcement and civil society come at this from different perspectives. The primary role law enforcement is, in essence, to protect the community and its citizens from harm and, in doing so, it must find a balance between security and privacy. (p.12)
• AI is described as a double-edged sword. It must be emphasised that AI and robotics are very much a double-edged sword, which can lead to great changes in the way in which law enforcement approaches policing or, just as easily, enhance the modus operandi of a criminal or terrorist group or even to create entirely new classes of crime altogether. (p.5)
• The report identifies three types of attacks: (1) digital attacks (cyber), (2) political attacks (media) and (3) physical attacks (drones).

Here is a list of cases where AI has been used straight from the report:

a. Autonomously research, analyze and respond to requests for international mutual legal assistance

b. Advanced virtual autopsy tools to help determine the cause of death

c. Autonomous robotic patrol systems

d. Forecasting where and what type of crimes are likely to occur (predictive policing and crime hotspot analytics) in order to optimize law enforcement resources

e. Computer vision software to identify stolen cars

f. Tools that identify vulnerably and exploited children

g. Behaviour detection tools to identify shoplifters

h. Fully autonomous tools to identify and fine online scammers

i. Crypto-based packet tracing tools enabling law enforcement to tackle security without invading privacy.

Four features of ethical development in AI Law Enforcement

I found that the report described that AI and Robotics in law enforcement should be characterised by four features:

Fairness: it should not breach rights, such as the right to due process, the presumption of innocence, the freedom of expression, and freedom from discrimination.

Accountability: a culture of accountability must be established at an institutional and organizational level.

Transparency: the path taken by the system to arrive at a certain conclusion or decision must not be a ‘black box’.

Explainability: the decisions and actions of a system must be comprehensible to human users.

Recommendations from the report

The report recommends actions in five areas.

(1) Research and analysis:

• Law enforcement needs for AI and robotics should be identified, structured, categorized and shared to facilitate the development of future projects.
• New or ongoing AI and robotics initiatives should be identified and mapped, with law enforcement agencies in the Member States being informed.
• The acceptable legal and ethical boundaries for data collection & analysis for and by law enforcement should be clarified.
• Opportunities and techniques for addressing privacy and accountability issues using AI should be investigated.
• A study on “new crimes” involving the malicious use of AI and robotics should be conducted.

(2) Awareness:

• Greater awareness of AI and robotics issues should be developed in law enforcement agencies through improved education and information exchange.
• The AI and robotics technology landscape should be continuously monitored.

(3) Meetings and coordination:

• A forum, such as the one created during the INTERPOL-UNICRI meeting, should be kept active, through annual meetings, to facilitate further discussions on AI and robotics for law enforcement and like-minded organizations.
• International cooperation between law enforcement agencies and other relevant stakeholders should be coordinated.
• More events (meetings, training courses, and workshops) should be organized, drilling down on specific topics such as surveillance and video analytics, predictive policing, unmanned aerial vehicle deployment, ethical collection and use of data in law enforcement, AI for counter-terrorism etc.

(4) Knowledge and information sharing:

• The transfer of knowledge and experiences regarding AI and robotics throughout the law enforcement community worldwide should be facilitated.
• Relations between law enforcement, academia, industry partners and civil society should be encouraged and fostered.
• A platform for mutual cooperation, collaborative work and the sharing of expertise should be developed.

(5) New tools:

• There should be a greater emphasis on pilot projects to develop and test AI and robotics tools for law enforcement agencies worldwide.
• National Central Bureau in INTERPOL Member Countries should be provided with AI tools.

I hope you found this helpful.

This is day 38of #500daysofAI.

What is #500daysofAI?
I am challenging myself to write and think about the topic of artificial intelligence for the next 500 days with the #500daysofAI. Learning together is the greatest joy so please give me feedback if you feel an article resonates.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI