Problem Statement

Develop a Police feedback system whereby all visitors’ experiences are taken into account and then real time feedback can be obtained.

Background of problem statement

Currently to get the feedback of functioning of police station, generally inspection or decoy operations are planned. Sometimes, calls to some complainants are made to understand the response given by the local police. So, there is need to develop more holistic real time information feedback system. Even the SOPs for model police station expects that regular feedback should be incorporated in the system.

Challenge involved in solving the problem

* Non digital mechanism
* Records of all visitors not included – no uniform mechanism to find who all visited police station.
* Inadequate manpower at police station or supervisory level for follow up/ monitoring on daily basis
* Multi-lingual society/ visitors and less digital literacy
Solution required
Design a
1- bulk SMS tool about experience eg – Happy, not happy
2-Auto chatbot which ask questions about experience and then collate response
3-Some online portal to record detailed response
4-Portal/ website/mobile app similar to Instagram whereby a user can post specific experience (good or bad) tagging specific police station/office and then collating location/thana specific
4-Any other idea… to get feedback Only sharing personal experience only
Expectation from the solution/tool
* User friendly
* Mobile friendly
* Cost effective – not more than 2-5K/thana expenditure
* Simple to use
* Regular auto data entry and compilation

Problem Statement

Co-creation - Using AI, AR to train situations like mob control and scenarios like

a. 1000s of crowd in riot like situation

b. Naxal operation

c. Attacks on mosques/temples/police stations

d. Idol immersion at Ganesh Festival, Durga festival, Tajia etc. Or Develop a game about such police scenarios Coordinate with HTC Vive/ Occulus Rift

Background of problem statement

We see that issues of mob handling and crowd control are on rise and need tactical handling. Mob can vary as agitating students, peasants, employees with different approaches. So, there is need to upgrade police training whereby with use AI to tackle this problem. Police generally use Mob control drills whereby physically some of policemen act as mob, but then since they are all policemen or civilian, gravity of training does not match the real life scenarios

Challenge involved in solving the problem

1. As of now AI/AR not used in this domain

2. Those initial models made look like video games – should be more realistic

Solution Required

Digitalize/ Augment Mob Control Drill used by police using Augmented Reality so that police personnel can have near real feel of dangers associated and linked risks while handling and dispersing those crowd and at the same time ensure own safety.

Expectation from the solution/tool

1. Multiple scenarios to cater to multiple scenarios

2. Multiple people to act in tandem in training as they will work/react/act in real life scenarios

3. Cost effective

4. Comfort level in wearing of glasses for near real feel of actual scenario

5. Should give a real feel in terms of sound, weather, wind etc.

Problem Statement

Leverage AI system response to create triggers/alarm when some suspicious activity is observed in lockup. Broad based use for camera analytics -

Background of problem statement

Custodial torture/custodial death is one of the most challenging and demeaning aspect whereby police excesses are alleged/reported. So, preventive strategy/ monitoring need to be develop.

Recently, in all police stations high resolution cameras are installed in police stations including facing police lockups. So, with regular information feed into AI system, we can set some triggers whereby some suspicious activities are observed in lockup – both by accused or the police.

Challenge involved in solving the problem

1. Differentiation between suspicious activities and normal activities will need multiple trail and error process

2. data set training – however such data set are not publicly available

3. Will be more complex with multiple people in lockup

Solution required

An automated system which will automatically alert the police sentry at the police station as well as the district police control room when some suspicious activity is observed by camera.

Expectation from the solution/tool

1. Should be able to give real time triggers

2. Should be able to identify risky objects like blade, metallic object, eatable etc.

3. Should be able to identify risky movements eg. Climbing on wall, drilling wall, taking something out from dress, wall etc.

Any other issue to be specified

There are 6-8 cameras installed in police station ( apart from lockup) – so can use them also for intelligent decision making Can also be used to track gang wars in jails (face tagging along with suspicious movement)

Problem Statement

Drafting a sound FIR (First Information System using AI and ML)

Background of problem statement

Whenever the complainant comes with a specific complaint – the FIR is registered after writing Police action with section of IPC, different acts as per the individual knowledge of IO and there is limitation in understanding of multiplicity of acts. Due to this there is allegation of application of wrong or weaker section of law against culprits. AI can help to suggest relevant sections as per the language of complaint and then IO can investigate accordingly.

Challenge involved in solving the problem

1. Datasets of CCTNS not being used for ML for this purpose (FIRs available on Rajcop )

2. Datasets available in local/regional language

Solution required

software/portal which will do OCR of written application and then process it to suggestion relevant section applicable as per given complaint.

Expectation from the solution/tool

1. Should be multi-lingual

2. Required to have OCR feature for major Indian languages ( at least Hindi for starting)

3. Should have feed local special laws along with CrPC, IPC

4. Should give the IO/Registering person to choose which al sections l to apply

Any other issue to be specified

Once major head of FIR is identified (eg. NDPS, crime against women) – can use open datasets including case laws, circulars, SOP for thorough detailed investigation and thereby improving the quality of investigation.

Problem Statement

Check authenticity of websites, phone numbers from Google through citizen complaint reviews , web scraping and other relevant methods

Background of problem statement

With increasing use of online shopping, there is rise of fake websites and Instagram ads, facebook ads etc. which claim cheap products but often delivery fake or poor quality products and at times even does not deliver. So, these websites and ads clearly cheat people.Similarly, many websites promote ponzi schemes or other cryptocurrency schemes which promote/guarantees 3-4X returns a year, which fall into unregulated banking deposits scheme and clearly criminal activity. Also on some google search for customer care number of some banks or Amazon/Flipkart instead of real number, number of a scamster figure in , whereby these fraudsters allure people and then do phishing to create money losses.

Challenge involved in solving the problem

1. Such websites are on the rise

2. Once one gets down, other with similar name is developed -so need for continuous search and monitoring of such websites

3. Multiple varieties including cheap shopping of cloth, shoes, sports equipments, job interviews.

4. Need to match up with changes made by websites and numbers

Solution required

A automated system which use publicly available data, consumer forum data, website reviews, web scraping, website analysis to identify all dubious websites, forums, advertisements or mobile apps and report that to police.

Expectation from the solution/tool

1. Should use all relevant information including publicly available data, consumer forum data, website reviews, web scraping, website analysis

2. URL ref flag, suspected domain history, pixelated images, typos and bad grammar, poor website design,bargain basement prices, limited contact details, complex-non existent return policy, questionable payment options, bad social media be used for identification of such suspicious websites

3. Should be able to fetch data from Cycaps portal (having repository of fake/ scam/ mobile numbers used in criminal activities and cyber crime)

Any other issue to be specified

Find about what I4C are doing.

Advisories

URL blocking as step further

Keywords selections

Developing crawler

Problem Statement

Geo-tagging of private cameras and trigger system about repairs along with identification of nearby camera location using app/website or portal Standardization of cctv camera, control and data protocols

Background of problem statement

Cameras are very crucial to track and investigation serious criminal cases. However it becomes very tedious for police to actually identify where all cameras are installed. More importantly, this creates a delay which is beneficial for the accused who thereby is able to run away using this delay. At times, it is observed that although cameras were installed but those are not functional or usable. So there is need to improve this scenario.

Challenge involved in solving the problem

1. Private cameras list not maintained

2. If some camera is ill maintained, gone wrong or some obstruction ( tree growth, darkness, chips packet, cloth, spray) used – no trigger sent

Solution required

Geo-tagging of private cameras & as per licensing capabilities to Command and Control centre + mechanism to check if a specific camera installed at particular location ( bus stand, toll, railway station, restaurant, petrol pump, shop ) has gone out of place and thereby automatically notifying that it needs repair

Expectation from the solution/tool

1. GPS tagging of each camera in city/locality

2. Area wise (in proximity feature) such as Hotels/places to visit near me by Google

3. Ability to check functionality of each camera linked and send trigger to the local police station and owner to repair/reorient it as per need.

4. Use of AI and ML to assess major crowd congregation which can hamper law and order situation or fight at restaurant and then sending this information to the local police station.

5. Feature to read/ process multiple video formats coming from variety of cam

Problem Statement

Design of a financial Data analysis software for financial frauds.

Background of problem statement

Cyber criminals by way of their numerous deceptive techniques successfully fool the complainant to transfer his/her money to certain bank account, ewallet, payment gateway or to any purchasing app like Amazon or Myntra. The amount once debited from the complainant‘s account flows through numerous transactions across the bank accounts, UPIs and payment gateways. Analysis of flow of money is very crucial to trace out the criminal. Generally, all the data is sought by investigating agency from the payment gateways. Every payment gateway delivers the information in its own format. Presently, only MS-Excel is used for analysis of data on a piecemeal basis. There is no integrated software where such data can be analysed.A dedicated software is needed which can compile the information received from different payment, gateways and banks in different formats and describe the flow of money in graphical format as well as the information of the owner of the bank account. Special attention needs to be paid to withdrawal of money either through online shopping platform or POS machine or ATM transaction

Challenge involved in solving the problem

1. Multiple datasets

2. Non- uniform data

3. Cross- integration of portals including payment portal, ATM bank records, bank data, CDR and locational analysis

Solution required

Design a software to analyse all the financial details made available by the financial institutions in an integrated way

Input:

1. Database of all banks

2. Their physical location on the map

3. SContact details of the nodal person for giving bank details

4. Contact details of the security incharge who can provide CCTV footage

5. ATM number corresponding, address and location on the map

5. POS machines given by the bank and their username

6. Details of the nodal person and the security in charge of all the financial institutions and payment gateways

7. File containing details of financial transaction which is made available by the concerned financial institution or payment gateway. These files can be in various formats like PDF Word, MS Excel or any other format.

8. Bank account owners information provided by the bank which includes his photograph, address, mobile number and other details. This information can also be in various formats

Expectation from the solution/tool

Process:

The software should first map the permanent details of banks, ATMs, POS machines in a unified format. Every transaction across the bank is uniquely identified by a transaction number. Using this transaction number, the input files of financial transactions needs to be mapped and results should be shown in both Excel format and the graphical presentation

Expected Outcome:

1. Both the presentation should simplify how the money has travelled across the payment gateways including details of time and place. The presentation should conclude at the point where money was withdrawn along with the geographical location of the ATM or the POS machine.

2. If any item has been bought through these transactions, then the address of delivery of such goods.

3. By clicking on any transaction, the details of the bank account as provided by the bank should be displayed including the photo of the account owner.

4. It should also show in the transaction summary of the bank account used

5. A fake bank account is generally marked by following features:

A. The KYC process from the bank is incomplete

>B. Money transaction doesn’t happen through the home branch

C. Nature of flow of money: money comes from various bank accounts and either withdrawal takes place or the money goes to one or two bank accounts.

6. Depending upon the above characteristics and AI based tool within the software, it may indicate that the probability of bank account being in fake or fictitious name is high.

7. Once the account is verified as a fake or fictitious account, the software should keep a database of such account so that if the same account appears in any other financial transaction, it should be able to map and identify the bank account across the transactions.

8. If multiple accounts are created in the same mobile number, it should be able to indicate the same.

9. Cdr analysis software are already available and police force is using them. Cdr analysis software gives the tower location and route chart of the suspected number on the map. An extended feature of this software can be to club the location data from the Cdr of the suspected person with the geographical location of the ATMs and POS machines used by the cyber criminal for withdrawal.

Any other issue to be specified

Should be able to leverage Cycaps and pan- India data about suspected mobile numbers and bank accounts.

Problem Statement

Develop a mechanism or system which can differentiate deep fake and originally created photos, videos and audio

Background of problem statement

With increasing usage of deep fake technology, deep fake is being used to spread disinformation, inflame religious, racial, cultural, ethnic tensions which can have huge ramifications on law and order as well as economy and polity of the country. So, there is a clear need to develop a technology to identify deep fakes and create alerts against such deep fake material in circulation.

Challenge involved in solving the problem

1. Multiple tools and techniques to develop deep fake (eg Faceswap, DeepfaceLab,StyleRig, Neural Voice Puppetry)
2. Multiple deep fake detection methods using Generative Adversarial Networks
3. Less understanding of police on these topics as these are recently emerging topics and need continuous upgradation of skills

Solution required

Develop a mechanism or system which can differentiate deep fake and originally created photos, videos and audio on upload/accessing open public data on real time basis

Expectation from the solution/tool

1. . A portal or website where suspected audio, videos and photos can be uploaded and through the developed model
2. Proactively access the public data and trending vidoes/audios/photos on Instagram, twitter to check their authenticity regarding deep fake

Any other issue to be specified

Refer - Deep Learning for Deepfakes Creation and Detection: A Survey – Aug 2022 research paper by Thanh Thi Nguyena, Quoc Viet Hung Nguyenb

Problem Statement

Develop a solution that simplifies the process of recovering money frozen due to cyber fraud for complainants.

Background of problem statement

The 1930 Cyber Fraud Helpline has been successful in helping victims freeze their money in cases of cyber fraud. However, the process of getting the money back is difficult for complainants due to the stringent provisions of Section 102 of the Criminal Procedure Code (CRPC) in Indian law. The 1930 Cyber Fraud Helpline is a helpful resource for victims of cyber frauds to report their cases and freeze the related funds

Challenge involved in solving the problem

1. Complex legal provisions: As mentioned in the problem statement, the recovery of frozen funds can be difficult for complainants due to the stringent provisions of Section 102 of the Criminal Procedure Code (CRPC) in Indian law. This can lead to a delay in the resolution of cases and cause additional stress for victims.

2. Limited reach: The 1930 Cyber Fraud Helpline is currently only available in a few cities in India, which limits its reach and effectiveness in helping victims across the country. This can be a disadvantage for people who do not live in these areas or are not aware of the helpline.

3. Lack of awareness: Despite the availability of the helpline, many people are not aware of its existence or how to use it. This can prevent victims from reporting cyber fraud cases and availing the services provided by the helpline.

4. Limited resources: The helpline may not have sufficient resources to handle the volume of cases reported, which can lead to delays and inadequate support for victims.

5. Complexity: The process of freezing and recovering funds can be complex, which can create distrust among victims and erode their confidence in the helpline.

Solution required

Develop a solution that simplifies the process of recovering money frozen due to cyber fraud for complainants post 1930 complaints and use of AI system for better decision making

Expectation from the solution/tool

1. Design an automated platform that streamlines the entire process of recovering frozen funds, making it easier for complainants to file claims and track the status of their cases.

2. Develop an AI-powered system that analyzes the details of the cyber fraud case and suggests the most appropriate legal steps that can be taken to ensure the recovery of the funds.

3.Provide legal guidance and support to complainants throughout the recovery process, simplifying the complicated legal provisions for them.

4.Implement a system of incentives and rewards for reporting cyber fraud cases and successfully recovering frozen funds, encouraging more people to come forward and making the process more efficient.

5.The solutions should take into consideration the legal framework and requirements of Section 102 of CRPC while also prioritizing the ease and convenience of the complainants.

Any other issue to be specified

A demo of current procedure of freezing, movement of money for frozen amount and legal provisions taken for send the frozen money to complainant will be provided.

Problem Statement

Design a Drone use for police purpose

Background of problem statement

Drone use has become inevitable in current times, effective usage can help in crowd assessment and crowd control. As of now drones are used used as cam recorders to understand the flow and density of crowd, but not give clear number of people and also not have function of usage of PA and non- lethal ammunition.

Challenge involved in solving the problem

1. Increased cost with adding features.
2. More inclusion of feature means more weight and thereby less fly time.
3. Optimization of battery and camera parameters

Solution required

Develop a drone – with public announcement, tear gas delivery with camera number estimation.

Expectation from the solution/tool

Develop a drone with capabilities like.
1. Public announcement system incorporated.
2. Camera to have crowd number estimation and then identifying mischievous characters also.
3. delivery of capsi-spray/ tear gas on command.

Any other issue to be specified

Can add other relevant feature to improve the functionality for police purpose.

Problem Statement

Dark web monitoring using keywords search with features of getting tor links, websites, geographical location

Background of problem statement

Dark Web is increasingly used by criminals for different activities and LEAs are least prepared and trained to tackle it. Thus it is much needed to develop a handy tool for Dark Web Intelligence which could help the police in tracking down the websites, gather intelligence and also trace the person, groups involved in any illegal activity.

Challenge involved in solving the problem

1. Accessing the darknet through TOR/I2P/ZeroNet.onion domains/.i2P domains.
2. Traffic through relays.
3. Less understanding of search engine tools for the dark web (eg. Katana, Onion search, Ahmia search engine, Darksearch).
4. Regularly evolving features and tools to get onion links from the darkweb (OnionScan, Onioff, Onion-nmap)and scrap Darkweb (TorBot, TorCrawl, OnionIngestor )

Solution required

The task is to develop a tool that could search for keywords across dark net search engines, along with an alerting mechanism.

Expectation from the solution/tool

1. must be a simple interface/Console which must connect to VPN and then help to find the details rather than police searching for websites, links, platforms on dark web.
2. It must also be able to fetch data from different browsers like TOR, TAILS, I2P, Freenet etc. Participants must develop a simple interface/console which automatically latched to a VPN and then connect through Dark Net Browser/Deep Net browser.
3. And then it should be able to filter data based upon Keyword(like Murder, Arms, Drugs,etc.), Geo-location, state, district, date, etc.
4. Extension: Expose IP details of dark net sites & users.
5. Tool which could identify a seller of any product and provide ways to track down the person. It should be able to able to gather/crawl details about all websites, chat rooms,shopping sites & other information related to dark web. It should be a handy intelligence and data collection tool on the dark web.
6. Web archive time-based proxy with easy User interface for access, Geolocation identification, TOR Keyword Search, Dark Web Crawlers

Any other issue to be specified

Explore multilple tools to create data collection architecture ( Scrapy, Tor, OnionScan, Privoxy, Elastic, Redis etc.)
Refer https://darknetdiaries.com/episode/

Problem Statement

Crypto-currency forensics and funds flow visualisation tool.

Background of problem statement

With the rise of cryptocurrency usage in fake dating profile (pig butchering) , ransomware, investment scams, darknet criminal enterprises, ponzi schemes and then investigation of cryptocurrency flow becomes difficult.

Challenge involved in solving the problem

1. Tracking the flow of cryptocurrency from hiding in mixers (tumblers), coin swap services, privacy coins, DeFI services, Layering services.
2. Limited features in free software likes Blockexplorer, Metasleuth, Etherscan3. Cross- border flow of money.

Solution required

Develop a better visualization software/ funds flow system of crypto-currency.

Expectation from the solution/tool

1. Can use features of Sankey diagram to show.
2. Incorporate features of what kind of hiding is done in flow ( mixers, DeFi services, Layering )
3. Direct approach whereby money flow to Crypto-exchange (eg coinbase, binance, kraken)
4. Integrate direct message to nodal of crypto-exchanges and mixer services to get details at the earliest.