Introduction
Our ever-changing environment requires creative thought since innovation allows us to continue to adapt to ongoing changes and to the unique demands brought on by digitalization. Psychologist J. P. Guilford asserts that creativity is a creative aptitude; an ability to create, in a complex integration of different skills that allow for the formulation, repackaging, and expression of various experiences for the first time in a new context (URL1).
Creativity implies imagination’s ability to create new ideas and to realize such ideas in a concrete form. To be creative means finding new ways of combining old thoughts or reorganizing past experiences in a new way to create new, original, and unique ideas in order to provide ingenious solutions to problems (URL2).
NASA scientists tested 1600 children aged 3-5 and then again around 10-15. As of the first age group, creativity levels tests manifested at 98%, whereas at age 10 they resulted only at 30% and 12% at age 15. Creativity levels were also tested on 280,000 adults with an average of only 2% of their creativity levels (Szlafkai, 2018).
Generative language models in artificial Intelligence are now more creative compared to humans in doing divergent reasoning tasks (Kent et al., 2024). Divergent thinking uses imagination and creative features to come up with an idea, by opposition to convergent thinking, which is logical and very direct. According to certain researches of 2024, 70% of employers believe that creative thinking was the most important skill (URL3).
The dynamic, fluid world of employment and criminology requires the utilization of human imagination. Imagination and creative thinking are integral for forensic (criminal) analysis. Without creativity innovation is not possible, and without innovation understanding historical events may be impossible.
Artificial intelligence can help us harness creativity that exists and we need to use it in the investigation. According to some researches, AI is 40 times faster than humans at generating ideas (URL4).
We must work with artificial intelligence as our new colleague. Consequently, mapping the potential of existing algorithms (AI) in solving issues, furthermore, analysis and introduction of methods (issues) necessary to effective use of AI is elemental, proving that AI can act creatively in performing certain tasks. Brainstorming with AI may facilitate task performance, and common idea-making may become a tool for effective work.
2. Research methodology and objectives of the study
Focusing on the role and the potential of artificial intelligence in both creative and problem-solving ventures with an emphasis on forensics, this paper will attempt to delve deeply into these technological developments that are capable of redefining law enforcement underpinning the importance of AI in assisting investigative ventures. The primary techniques employed in the study are the five whys method (to reveal roots), the Kipling method (5W1H questioning technique), fundamental forensic questions, AI vs human performance pilot study (with 100 individuals to compare human and AI creativity), and individual experiments in practical communication with AI in a tangible investigative case.
3. Problem-solving techniques
Albert Einstein is noted for having said, “…if I had an hour to solve a problem, I would spend 55 minutes thinking about it and 5 minutes solving it” (URL5). Indeed, Einstein stressed the importance of understanding and knowing a problem to the end of its firm roots and determining what it is all about. By understanding these root causes, he says, we are not only armed to take on the issue at hand but we are also capable of effectively solving it. This human-computer and computer-artificial intelligence interface and AI ideas can result in the final comprehension of the core problems and, therefore, they may also assist in solving the problems. Still, co-creative thinking in the course of successfully execute the desired task requires detailed formulation of the problem with AI. This is a major development since it would guideline and direct artificial intelligence to provide a structure for the correct type of action among many alternative possibilities, showing what particular patterns or patterns would be applied based on a large database.
3.1. According to the 5 Whys technique
The root cause can only be understood with the help of deep research. If one gets questioned repeatedly five times in a row about the cause behind an event or happening, experience proves that the root cause will be successfully identified, nevertheless, it can also be determined in fewer questions (URL6).
If you repeat the why question five times (generally, five why questions are sufficient), you reach the root cause, or more accurately, to the “true” root cause. The responses may involve aspects we had not considered so we can reach to the actual cause (URL7).
Several academic analyses and studies quote and refer to the case of permanent deterioration of Washington DC’s Jefferson Memorial through it must first be noted that its structural condition has only recently deteriorated. The reason, or to be particular, the root cause, is explained using the 5-W’s in this way:
Why1 – Why is the monument breaking down?
– Because people used to clean the monument using harsh chemicals.
Why2 – What is the need for harsh chemicals?
– To get the excessive amount of bird droppings removed from the monument.
Why3 – Why do bird droppings accumulate excessively on the monument?
– Because the monument has a large congregation of spiders around it, which attract local birds to feed on them.
Why4 – Why do many spiders gather near the monument?
– Because there are large clouds of flying insects there in the evening, on which spiders hunt. Why5 – Why do the same clouds of insects form near the monument in the evenings?
– Because evening illumination of the monument attracts local insects.
Solution: Change Jefferson’s dusky lighting to something that does not attract swarming insects. The fix was really simple and even saved the Parks Department money, namely, just turning on the lights after dark (URL7).
The story of the Jefferson Memorial is often used in the context of the 5W method, although we must also stress here that this apparent transparency of the method should not deceive us, as the matter of problem-solving is not necessarily easy to do (URL7). Actually, the 5W approach is just one of the possible methods, and in my opinion, it does not always leads to the root cause, however, in terms of AI, it is surely one of the potential brainstorming measures, hence the approach of 5Ws even if it would not result in an actual outcome, every response might give an idea or version that might assist in, for instance, the exploration of a historical event facilitating forensics to understand what had happened.
3.2. The Kipling Approach
5W1H also is a problem-solving questioning method. It is an acronym of 6 English question words for gathering information (Who, What, Where, When, Why, How). It is one of the simplest methods of exploring problems. Gathering facts may only be effective if we ask the right and appropriate questions (URL8).
An example of the 5W1H method: if a patient complains of a headache, you are able to arrive at the root cause by using the questions of the Kipling method (URL8):
– What does the patient have? – A headache?
– Whose head hurts? – John Doe
– When does the head hurt? – In the morning
– Where is the headache? – At the temples
– How does the head hurt? – Throbbing
– Why does your head hurt? – High blood pressure
The symptom vanishes as soon as antihypertensive agents begin to number down blood pressure, but the underlying cause persists, again demonstrating that the why question might give the solution to the real root cause:
– Why is your blood pressure high? – Because his blood cholesterol level is too high. (2. Why)
– Why is your cholesterol high? – Because he consumes too much fat (3. Why)
This brings us to the basic underlying problem (URL8). This is where an element of the 5W1H method (sometimes called the Kipling approach) is combined with the 5W method, where additional questions about the reasons can help to uncover the underlying cause.
The more data is fed into AI, the more specific it gets at the recommendations churned out. Moreover, always remember that AI is to be thought of as a collaborator, contributing a huge amount of ideas but never giving you that single definite answer.
3.3. Forensic thinking is the basic questions
Creative forensic reasoning is a permanent part of the investigation, in which the goal is to be gleaned by members of authority from the objective by mental reconstruction of past events and the use of causal relationships in the process and planning of the investigation. Both fantasy and creativity two of the most substantial elements of forensic thinking. Fundamental questions of forensics are also key instruments of the context of criminal thought. The forensic form questions provide a way-finding path so that the relevant incident can be understood within the framework of the past, present, or even future events. The available body of scientific literature shows variations in both the number and format of the fundamental form questions. The well-known types are: what happened, where, when, how, who, with whom, and why. Generally, the basic issues of forensic science can be enriched by talking about the interrelation of events, at the same time, considering the plan and method of action.
It is key to find a way of communication, “cognition”, and more to the point, to work with artificial intelligence. To be sure, certain aspects of forensic approach, insofar as they involve imagination, are very pertinent to the discussion here. The essence of creativity essentially comes into play when there is a need to procure collective intelligence. Still, in the evaluative act, there is a sense of scepticism should also be taken into consideration as an element of the control mechanism. Artificial intelligence (AI) should first and foremost be viewed as an investigative tool, in more or less the same way a new investigative assistant or partner would support collective intelligence and forensic thinking through the process of brainstorming.
To sum it all up in simple terms, communication with artificial intelligence is a necessity event if it observes all prompts and protocols. In the context of using artificial Intelligence within the scope of “collective conversations”, the following techniques are significant: the Kipling technique, that is, the 5W1H questions, the list of 5W questions, especially the five why questions, and the basic criminological questions. The basic questions within criminology are very similar to those asked by the internet clipping method; however, a major difference is that the basic criminological questions are supported by the additional question of “with whom?” specifically focusing on criminal investigations.
By asking questions at each point of the communication process, we facilitate artificial intelligence’s better understanding of exactly what is being requested from the vast amount of data that it has access to. It is often more effective to encourage the AI to express its train of thoughts aloud verbally. This process enables me to catch moments where AI has misunderstood some information within the course of conversation so I can intervene and make any crucial corrections during the process if needed. Within this technique of verbal communicating with the AI, I can have AI to explain how it has come to a decision or why it answered the way it did. Then, it will lead me through all of its logical turning points of the reasoning process, revealing its nature one could describe as a “massive resource loaded with much information”. Verbal reasoning allows you to better understand the context. In summary, the ability to articulate AI reasoning is useful because it:
- makes the AI decision-making clear as well as the logical background behind the process,
- helps us understand how the AI got where it got,
- makes it easy to spot errors/inaccuracies,
- allows us to learn from AI’s problem-solving methods.
By unlocking new dimensions, cooperating with AI is essential in the forensic field.
4. Creative thinking with AI in an experimental exploration
The experiment was prepared in order to understand the relative creativity of human versus artificial intelligence within a specified time restraint. The main hypothesis indicates that artificial intelligence creativity would be greater than human within a specified, limited time.
The AI, in addition to the 100 original interview participants, was tasked with using their imagination to enlist the possible ways of using a pin. They all had 1 minute to complete this.
Within 2 seconds, the AI was able to effectively complete the task, identifying 6 main categories and 8 possible ways of use within each category. Overall, within that 2-second time limit, a total of 56 potential uses were discovered.
I put this question to Claude ai during the experiment, requesting it to list as many potential cases of use as possible for a pin. The answer from Claude ai:
“The uses of the dowel pin:
Crafts and decoration:
- Jewellery making
- Embroidery, brooch making
- ’Embroidery and accessories’, ’Embroidery for jewelry’, ’Embroidery for accessories’
- Making Christmas tree decorations
- Making paper decorations
- Stringing beads
- Temporary fixing of ribbons
- Shaping textile flowers
Clothing and textiles:
- Fixing patterns to fabric
- Preparing garment repairs
- Temporary fixing of hems
- Positioning of collars, laces
- Fitting of garment linings
- Marking the fold of curtains
- Decorating hats
- Fixing the edges of carpets
Education and art:
- Map markings
- Fixing presentation materials
- Wall hanging of sketches
- Assembling models
- Stretching canvas
- Positioning drawings
- Fixing exhibition pieces
- Making school tables
Scientific and professional:
- Insect art collections
- Acupuncture treatments
- Electronic testing
- Microscope specimen preparation
- Marking of anatomical models
- Cartographic markings
- Surgical instrument (sterile)
Precision measurements
Household and everyday:
- Paper crumpling
- Fixing calendars to walls
- Temporary display of photos
- Hanging curtains
- Fixing mosquito nets
- Fixing Christmas lights
- Supporting plants
- Used as a safety pin
Special applications:
- Fixing fishing baits
- Hole punching in modeling
- Closure (not recommended)
- For use in fishing netting (not for fishing netting)
- Assembling costumes
- Repairing musical instruments
- Minor assembly work
- Making temporary markings.”
While the age of persons (respondents) was variable, their gender was also mixed (male, female). Respondents mostly identified the following categories: pinning things on walls, stapling papers, marking on maps or blackboards, making hats and various objects, making clothes, decorating, making signs on blackboards, making eye-catchers, combing, and decorating hair. Respondent’s breakdown (n = 100): in 1 minute, four responses (60 people) (60%); 3 responses (27 people) (27%); 2 responses (10 people) (10%); 1 response (3 people) (3%).
60 respondents identified 4 possible uses, 27 respondents identified 3 possible uses, 10 respondents identified two possible uses, and 3 respondents identified 2 possible uses (see Figure 1).
Figure 1.
The hidden motive of the experiment was to identify the distinction between AI and human creativity in enumerating potential uses of the pin. The AI provided 56 potential uses in just 2 seconds, whereas human respondents needed at least 1 minute of times to note potential uses, and they provided a total of 4, 3, 2, or 1 means of possible use.
AI performance:
Time: 2 seconds
Possible uses: 56
Human performance:
60 individuals: 4 potential uses.
27 individuals: 3 potential uses
10 individuals: 2 potential uses
2. Figure
According to the above, AI would provide 56 potential uses within 2 seconds, while human interviewees enumerated a minor number of possible uses within at least 1 minute. 60 individuals identified 4, 27 individuals 3, 10 individuals 2, and 3 individuals two potential uses.
The experiment upheld our hypothesis, through which we postulated that AI could yield creativity better than a human. The test clearly demonstrated that AI progresses and produces creative ideas more quickly and to a greater extent compared to humans.
5. Joint criminal work with AI
5.1. Investigation plan
During an investigation, the criminal rapporteur works mostly independently and undertakes a number of investigative measures. Early in the course of the investigation, the criminal investigator develops an investigation plan in the format of a conceptual framework, a memo, or a written summary. The conceptual framework generally remains at a theoretical level. A mental plan, especially considering the probable loss of information, is not preferred. What the investigator or interrogator may forget are the questions considered relevant and necessary under pressure from externalities. From a forensic perspective, note-taking is the preferred approach. Note-taking would be an informal writing down of the planned actions and their time frames by the seller of the case, not bound by procedural law, as opposed to a written plan with the proper procedure described. Pursuant to Paragraph (1) of Article 134 of Government Decree No 100/2018 (VI. 8.) on the Detailed Rules of Investigation and Preparatory Procedure, the authority conducting the investigation shall draw up a written plan for the investigation,
a) when there are difficult facts and/or legal findings or
b) in the establishment of a joint investigation team or
c) as a result of a decision made by an authorized head of authority conducting the investigation.
(2) The investigation plan shall include
a) a summary of the historical facts of the criminal offense under investigation in relation to the statutory elements of the offense and the classification of the offense in the Criminal Code,
b) description of the scope, context, and detail level of the facts to be ascertained,
(c) the procedural steps to be taken in order to ascertain the facts,
(d) the objective to be achieved by the procedural act,
(e) the person responsible for carrying out the procedural steps,
(f) the time limit for the performance of the procedural steps,
(g) the measures envisaged for the recovery of the property, the procedural steps to be taken, and
(h) the disguised means to be used.
(3) The investigating authority shall reconsider the investigation plan in accordance with the state of the proceedings and, where warranted, make any changes or additions thereto.
Apart from the available data, the preparation of an investigation plan is, to a great extent, influenced by the experience and logic of the investigator, knowledge of the type of case, the investigator’s imagination, and finally by elements of forensic thinking from the beginning till the end. With advancements in technology, AI can act as a collaborator or member of an investigative team to complement and stimulate the ability of the investigator. When working together, a detective can come up with a set of versions.
These versions should also be indicated in the plan. In addition to the above points, the investigation plan should indicate the possibilities and means of using artificial intelligence. The investigation plan should be updated if, for example, existing data are checked or new data emerge.
5.2. Communicating and brainstorming with AI
For collaboration and brainstorming with AI to be successful, prompts need to be developed. However, more effective results would be achieved by including context and providing continuous information exchange and text input with the AI.
Prompts would typically be understood as text-based commands or instructions, but surely, voice-based prompts will soon be a trend. Input or instruction, which is used to add it to the language model to elicit the best response, can be termed as prompt (Bereczki, 2024).
However, communication with generative AI can be much more effective than a well-written prompt, as the following conversation with AI demonstrates.
The next assignment was no longer given to Claude ai but to another AI program known as Chatgtp. The author gave AI the following historical facts:
“At the university, the lecturers and the students are working on a crime case. The aim is for students to draw up a very detailed investigation plan, in which they indicate the specific tasks they want to carry out, with a planned time for their execution. The available data are as follows: On 10 December 2024, an unknown perpetrator went to a property in Budapest in order to break in and steal valuables in the absence of the occupants. The offender approached the property from the direction of the neighbouring plot, then broke down the wire fence separating the plots and pushed in the window left open. Once inside, the man stole gold, silver, and biscuit jewelry and a laptop computer. A few months later, on the evening of 07 January 2025, the offender targeted a condominium in Budapest, broke down the fence, and then boarded up the terrace of one of the ground-floor apartments. Once inside, he stole cash, a mobile phone, and silver jewelry. Furthermore, according to the available information, the perpetrator could not be identified, but there was camera footage of the perpetrator at both locations, which was shown on television news in the evening. After the news aired, police received an email from a resident who described that he would recognize the offender.”
Key pieces of information given to AI in this one were academics, students, processing a crime, and a detailed investigation plan. It means that keywords would allow AI to identify precisely what one is asking it to help him with. So it enabled AI to select a particular selection pattern and search directions within its huge database.
AI’s answer was:
“Investigation plan
1. Systematisation and analysis of basic information (8-10 January 2025)
Goal: Structured processing of existing information to identify the perpetrator.
Tasks:
Analysis of camera footage:
– Detailed analysis of the perpetrator’s facial features and body type (e.g., height, build, clothing, tattoos, or other features).
– Motion analysis (e.g., gait, movement dynamics).
– Identification of objects, tools, or implements used by the offender.
Listing stolen items:
– Exact value of stolen items and unique identifying characteristics (e.g., laptop serial number, unique identifying characteristics of jewelry).
– Possibilities for selling stolen items online or offline (e.g., pawnshops, classifieds).
Historical context analysis:
– Similarities in the offender’s actions (e.g., entry methods, times, locations).
Tools:
– The facial recognition system is based on camera footage.
– Retrieving data from a network of CCTV cameras to map the offender’s movements.
2. Detailed analysis of field traces (11-15 January 2025)
Aim: To identify evidence of the presence of the perpetrator.
Tasks:
First scene:
– Examination of footprints (e.g., shoe size, type) from trampling the wire fence.
– Identification of fingerprints or glove prints left on the window frame.
– Collecting DNA samples (e.g., hair, skin fragments, sweat) left in the interior.
Second location:
– Identification of tool marks from fence demolition (tool type, characteristics).
– Identification of fingerprints and tool marks left on a patio door.
– Collection of DNA traces left in the interior.
Tools:
– Footprint analysis tools.
– Fingerprint analysis kit.
– Tool-print comparison with criminal databases.
3. Eyewitness interview (16 January 2025)
Purpose: To gather further information on the offender.
Tasks:
– Detailed interview of the reporting witness:
– Cross-check the information provided by the eyewitness with camera footage and other evidence.
4. Investigation of notifications from the public (17–20 January 2025)
Purpose: To gather information on the movement or whereabouts of the offender.
Tasks:
– Sorting and sorting of reports (based on credibility).
– Field verification of relevant tips.
5. Operational investigative steps (21–27 January 2025).
Goal: Locate and observe the perpetrator.
Tasks:
– Conduct surveillance operations at the suspected locations of the perpetrator.
– Search for stolen property (e.g., pawn shops).
6. Arrest and search (28–31 January 2025).
Purpose: To apprehend the offender and collect evidence.
Tasks:
– Preparation of the arrest (e.g., mobilization of an operational unit).
– Conduct a search to find stolen objects and tools.
Expected results and next steps
– Identification and apprehension of the offender.
– Partial or total recovery of stolen valuables.
– Bringing the offender to justice. “biography” (Chatgtp.com)
As requested, the AI drew up an investigation plan of extreme detail, including all tasks that had to be executed and the time they should take. Six major topics were found, and further areas of expectations for results and actions were broken down by the end of the investigation plan. It can be said that brainstorming with AI brings many aspects to the investigator’s attention. It is also possible to further analyze the relevant sub-points with AI. The above plan prepared by the AI can be considered as a so-called ’investigation-wide plan’. However, a sub-plan can also be prepared for the points contained in the plan, which will serve to improve efficiency and effectiveness at that level.
According to the data found, a witness expressed his testimony through an email, which is the most important of all the others, as the perpetrator’s identity was is still unknown to the authorities. As the identity of the perpetrator is revealed further practical investigative measures can be taken shortly. This may yield not only to more accurate and detailed understanding, identification and proof of the historic event but also to curtailing of necessary coercive acts (e.g., arrest or restriction of personal freedom), which, in turn, will be helpful to prevent the criminal perpetrator to engage in further criminal offenses. The testimony of the eye-witness is a crucial part of the investigation. Such testimony most probably will be the moving force of the investigation, therefore, the questioner must take a detailed and comprehensive statement as it is required and could be used as evidence later. Consequently, the detective should have in mind the technique of visualization and the technique of imagination that was elicited by the interrogator, so that the other person can ask the correct questions in the context of the witness interview. Thus, it is also necessary to investigate what questions AI would pose to the witness on the factors available. This is definitely necessary so that AI may determine the prompt questions for the interrogation process that the investigator might not have thought of before.
Given this, I asked the AI system the following question: “In the inquiry plan, you have mentioned the interrogation of the eye witness; what sort of questions would you ask him? I expect to have a list of questions in the most detailed way.”
AI’s response:
“Compiling a detailed list of questions is key to interviewing a witness, as the information they provide can help identify and apprehend the perpetrator. Below, I present a set of questions that are structured in a logical manner to maximize relevant information.
I. Perceptions about the offender
1. When did you first see the offender?
– Give an exact time, date, or time of day.
2. Where did you see the offender?
– Location of the sighting (street, building, area).
3. Under what circumstances did you see him?
– What activity was the offender engaged in (e.g., walking, running, standing still)?
4. How long did the sighting last?
– Duration of the sighting (seconds, minutes).
5. How many times did you see the offender?
– Was it a single encounter or multiple encounters?
II. Description of the offender’s appearance and clothing
6. What was the offender’s physical appearance?
– Height, weight, body type (e.g., thin, muscular, stout).
7. What was the offender’s facial appearance?
– Facial features: e.g., round, oval, angular.
– Special features: scars, scars, birthmarks, tattoos.
8. Hair colour and style?
– Short, long, bald hair colour (blonde, brown, black, etc.).
9. What was the skin colour of the perpetrator?
– Light, brown, dark shade.
10. What was the offender wearing?
– Outer clothing (jacket, sweater, shirt colour and style).
– Trousers, shoes, or other accessories (e.g., hats, gloves).
11. Was the offender carrying any objects?
– Bag, tool, box, or other visible object.
III. Behaviour of the offender
12. What were the offender’s movements?
– Did he walk normally, hurry, run, or act suspiciously?
13. Did the offender have any tools or equipment?
– If so, what type of tool did you see? (e.g., crowbar, hammer).
14. Did you see the offender talking to anyone?
– Was he alone or communicating with other persons?
15. How did the offender behave?
– Was he nervous, confident, in a hurry, or did he reflect other emotions?
IV. Previous experience
16. Do you know the offender?
– (personal acquaintance, sight, neighbourhood, etc.)
17. Have you seen the offender in the neighbourhood before?
If yes, when and where?
18. Have you heard anything from others about the offender?
– Possible rumours and sightings of other witnesses.
V. Further information about the offender
19. Do you have any idea where the offender might live?
– Indicate neighbourhood, street, or part of town.
20. Do you know if the offender is related to anyone?
– Name friends, family members, associates.
21. Have you seen the offender in any vehicles?
– If so, what type and colour of vehicle? Was there a registration number?
VI. Questions for camera recordings
22. Have you seen the camera footage shown in the news?
– If so, do you recognize the perpetrator?
23. What can you say about the movements and behaviour seen on the camera?
– Is it comparable to what you saw in person?
VII Other questions
24. Is there anything else you would like to share about the offender?
– Possible further observations and personal opinions.
25. Are you willing to further assist the investigation?
– For example, to act as a witness or to provide further information.
Comment:
The questions can be used flexibly, and the interrogator can ask for more in-depth details if necessary. During the interview, it is important that the witness feels comfortable and does not feel under pre” (Chatgtp.com)
The AI has built the interrogation plan by creating seven main topics and further detailed each issue by establishing 25 main questions. From my perspective, AI has developed a very ingenious and shrewd interrogation plan that has the potential to inspire the interrogator. It is possible that the AI may pose a question that the interrogator would not have even come up with.
I was concerned about further directions of communication with the AI, so I asked the following question: “If the eyewitness recognized the perpetrator, what kind of investigative act could be carried out?” I have received a rather specific response from the AI answer to my query that was as follows:
“If the witness recognizes the perpetrator, the following investigative actions become necessary and possible:
1. Personal identification procedure for the purpose of identification (presentation for recognition)
Its essence: The eyewitness is presented with photographs or persons that may include the perpetrator.
Steps:
1. Preparation:
– The police collects photos or personal characteristics of the perpetrator (e.g., taking into account the similarity of age, body shape, and height).
– The details of the procedure are discussed with the witness in advance, paying attention to ensuring objectivity.
2. Presentation for recognition:
– Several persons or photographs are presented to the witness, and they are asked if they recognize the perpetrator.
3. Taking minutes:
– The result of the identification is recorded in a protocol.
2. Operational monitoring and surveillance
Its essence: If the witness provides accurate information about the offender’s current location or habits, the police can launch a surveillance operation to locate the offender.
Steps:
1. Observation of potential places of residence (e.g., address, workplace).
2. Tracking and documenting the offender’s movements.
3. Discovery of other relationships (e.g., accomplices, helpers).
3. Production and questioning of suspects
Its essence: If the witness clearly recognizes the perpetrator, and additional evidence supports the commission of the crime, the authorities can produce the perpetrator.
Steps:
1. Production:
– The perpetrator is apprehended and taken to the police station.
2. Suspect interrogation:
– The perpetrator is informed about his rights and the fact of suspicion.
– The offender may make a statement, which is recorded in the minutes.
House search
Its essence: The perpetrator’s home or other property used by the perpetrator is searched for stolen items and the means of the crime.
Steps:
1. Obtaining a house search permit:
– Based on the permission of the prosecutor’s office or the court (in urgent cases, it can also be done without judicial approval).
2. Implementation:
– The police will search the scene and seize items related to the crime (e.g., stolen valuables, tools, clothing).
3. Documentation of evidence:
– The evidence found at the scene is recorded and sent for laboratory testing.
5. Confrontation
Its essence: The witness and the perpetrator meet each other directly as part of a police procedure so that the witness can repeatedly identify the perpetrator.
Steps:
1. He guides the perpetrator and the eyewitness through the details of the events with investigative questions.
2. The eyewitness can repeatedly confirm his statements about the perpetrator.
6. Comparison of other evidence
The description given by the eyewitness is compared with other evidence, such as:
– Camera recordings.
– On-site DNA traces and fingerprints.
– Items seized from the offender.
These actions help to identify and apprehend the perpetrator and prove the crime.” (Chatgtp.com)
According to section (1) of Article 210 of Act XC of 2017 on Criminal Procedures stipulates that the respective body, i.e., the court, the public prosecutor’s office, or the investigative authority, as the case may be, on the basis of the necessity of the recognition of the person or object, shall work out a statement for recognition. The defendant or the witness must be faced with at least three persons or objects for the purpose of identification. If none of these is feasible, a person or an object can be made to be seen and recognized by the accused or the witness with the help of a photograph, the equivalent of a sound, or a picture and sound recording. That is why I have asked AI the following question: ’How many people do you think the witness should be present during the identification process?’
AI’s response:
“During the presentation for identification, the witness must present at least four to six persons or photographs, including the perpetrator, if he has already been identified. The purpose of the procedure is for the witness to clearly recognize the perpetrator, while the possibility of a random or forced choice is ruled out. The presentation must follow some important rules:
Rules and Guidelines:
1. Presentation of persons with a similar appearance:
– The persons presented next to the perpetrator must have similar physical characteristics (e.g. height, body shape, hair color, facial features) in order to avoid intentional or subconscious influence on the witness.
2. Minimum number:
– For photographic identification:
At least 6-8 photographs must be presented, only one of which is the perpetrator.
– For personal recognition:
At least 4-6 people are needed, among whom the perpetrator may be present.
3. Perpetrator’s place in the line:
– The offender should be placed randomly in the presented group so that he does not stand out or it is not clear who he is.
4. Witness information:
– The witness must be warned in advance that the perpetrator may not be among the persons presented. This prevents the witness from being forced to choose someone.
5. Ensuring objectivity:
– The police officer performing the identification must refrain from any comments that could influence the witness’s decision.
6. Documentation:
– The procedure must be thoroughly documented, and a protocol must be recorded in which the choice of the witness and the justification are recorded.
Summary:
Increasing the number of persons presented improves the reliability of the procedure, so it is common to present 6-8 persons or photographs. However, a minimum of four people or photos is still legally acceptable. The basic principle of the procedure is objectivity and the complete avoidance of influencing the witness.” (chatgtp.com)
The Hungarian law requires at least three people to be presented. According to the information above, AI suggests a minimum of 4 people, that I personally find acceptable. It is evident that its is striving to get the most out of thousands of source patterns. However, it is the task of the interrogator to compare the recommendations proposed by the AI with the applicable law, as procedural legislation may even differ from country to country.
However, the goal was to emphasize that it is not recommended at all or, in general, to give directions (prompts) to the AI. On the contrary, it is more advisable to converse with it. If we talk to AI, it can facilitate the improvement of effective collaboration to a greater extent. In this way, it can be said that well-formulated questions and a continuous conversation enable the AI to get to know better what we are interested in, so it can give more accurate and relevant answers that can contribute to and facilitate the implementation of the investigative act, in this case, the interrogation and presentation for recognition.
First and last, it can be stated that the AI-generated questions and the investigative actions can be useful. The value and usability of AI has come out very distinctively in preparing the investigation plan and compiling questions to interview witnesses. The AI-generated questions were logically structured and were indicated to gather as much relevant information from the witnesses as possible. They included seven main topics, like the description of the offender’s appearance, his behaviour and movements, comments on camera footage taken, and other related comments. The AI-proposed questions were not only detailed but possibly creative to inspire the investigators by raising aspects they had not previously considered. This creative approach added to increasing the effectiveness of the investigation. Fort he case if the eyewitness recognizes the perpetrator, AI demonstrates in detail what investigative steps and actions should be performed, for example, presentation for identification, operative surveillance, production and questioning of the suspect, as well as the house search and its process. The procedures proposed by AI not only increased the efficiency of the investigation but also ensured objectivity and legality. This can bring a number of advantages in the cooperation between the investigative authority and the AI since AI is capable of analyzing massive amounts of data in a short period of time and finding correlations that could have remained obscure for humans. Besides, the answers proposed by AI can be of great inspiration to investigators, engendering new ideas. In such continuous dialogue, AI better perceives what it is required to do and, hence, provides more accurate responses. This is particularly relevant in areas such as investigation, where proper interpretation of details and a creative mind are of great importance.
6. Summary
Creativity is a decisive factor in solving the challenges of the 21st century, especially int he contexts of digitalization and innovation. Working together with AI in the framework of the experiment – formulating questions during the investigation – interprets how cooperation with AI can support investigative processes, especially detailed planning, interviewing witnesses, and identifying perpetrators. It may not only be be powerful tool but also a source of inspiration that raises new perspectives. Nevertheless, the user must always check whether the solution offered by AI are legal and relevant.
With such an application, artificial Intelligence also helps in enhancing the speed and nature of creative processes; it helps to enhance and understand problems that have been known for ages but were never actually solved. Cooperation between human and artificial itelligence is expected to open new perspectives of development in scientific research int he field of forensics and in many other sciences.
Proper communication is the basis of effective cooperation with AI. This primarily involves the formulation of precise and clear instructions, officially referred to as “prompts”. A prompt is a text command that specifies the direction in which the AI should start working. However, experience shows that continuous dialogue with AI can be much more effective than a one-time, well-formulated instruction. The conversation allows the AI to better understand the user’s intentions and provide more relevant, accurate responses.
The joint work of the interrogator (investigator) and AI opens a new era in problem-solving and creative processes, as AI is not only a tool but a partner that can complement detective skills and provide new perspectives. The relationship between AI and the investigator can, therefore, not only be useful but also essential in solving the challenges of the future and in the execution of investigative tasks.
Subsequently, the need for investigative authority to resort to AI is crucial, however, in order to make the best use of it requires the proper user skills. Presumably, this new modern potential will be familiar to more and people users in the future. There is a strict requirement for authority members to protect the interests of the investigation and in addition, to comply with the applicable legal statutory provisions and data protection regulations.
This study does not cover the potential risks of using AI, such as ethical issues or errors generated by AI, as this would merit the frames of a separate research. AI is part of the development of our history and, therefore, requires proper knowledge of its use. In addition to numerous positive returns, its negative effects must also be examined, as criminals may also take advantage of the opportunities provided by AI.
AI plays a pioneering role in the field of innovation and its use appears in almost every sector, from healthcare to industry. Current experience shows that if artificial intelligence is used for legal tasks and purposes, it speeds up work processes and reduces working time, as well as it reduces costs and expenses.
In the near future, AI assistants are expected not only to respond but also to perform tasks independently. What this will mean for the field of forensics is that by the time the investigator receives a criminal case, the AI assistant will have already conducted some kind of analysis in particular and will have presented certain recommendations – for example, what investigative actions are needed to be pursued. The final decision and responsibility shall rest with the investigator because the recommendations and versions are only put forward by the AI assistant that assists the investigation. It is also an important aspect that (both the investigator and) the AI assistant will remember the previous conversations, so joint ideation is expected to be even faster and more flexible. I believe a Police AI strategy should be developed, laying down the foundations of innovative forensic thinking. AI will also be an important participant for the investigative authorities in the future, and we can also see it as a good friend who doesn’t get tired and is always available. We can say without hesitation that classical law enforcement training, in the sense of criminology and forensics, has become a thing of the past.
7. Findings, values
It is much faster and more prolific than a human mind in devising creative ideas. The experiment above listed 56 possible uses in 2 seconds, while humans listed a maximum of 4 in 1 minute.
Continuous dialogue is better than one-time prompts when communicating with AI.
The AI can develop detailed and structured investigative plans and interrogation strategies.
The collaboration of human creativity and AI might open new dimensions in investigative processes.
Continuous monitoring of the state of the law regarding legal compliance and data protection is needed when applying AI.
The value of the study is that, based on concrete examples, it presents how AI and human expertise can create a synergy in the practice of law enforcement and offer practical guidance on how to cooperate effectively with AI. The research highlights the importance of the use of AI in modern investigative work while also drawing attention to the need to comply with legal and professional limits. The additional value of the study is that it supports the creativity capabilities of AI with empirical data and suggests concrete methods for effective communication with AI. However, it should be emphasized that the results of the experiment cannot necessarily be generalized to all creative tasks.
It is essential that we critically and thoroughly check the information generated by artificial intelligence. Investigators should be aware that the complexity of the algorithms can make it difficult even for experts to understand the analysis processes accurately. The results derived from the AI system can be termed questionable when the AI is not able to explicitly explain how it has generated a particular result. This issue is particularly relevant in relations with the matters of forensic examinations. In such examinations, findings produced by the AI during the inquiry cannot be considered as absolute truths. The investigating authority must treat whatever information artificial intelligence gives only as a possible version of truth until it has been proved genuine. Artificial intelligence has now become an immensely valuable tool that brings new opportunities for different industries, whether business, healthcare or research applications. However, in order to derive such opportunities to the maximum AI results should be evaluated under a critical and analytical approach, with special consideration to the operations of its algorithms, the reliability of data used, and their possible biases. Only then shall we ensure that the information generated by AI serves as true and useful grounds for decision making in order to achieve the objected goals.
1 A Kulturális és Innovációs Minisztérium EKÖP-24-4-II-33 kódszámú Egyetemi Kutatói Ösztöndíj Program – a Nemzeti Kutatási, Fejlesztési és Innovációs Alapból finanszírozott szakmai támogatásával készült. Supported by the EKOP-24-4-II-33 University Research Scholarship Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund.