Chatbots, like ChatGPT and Google Bard, are powered by artificial intelligence (AI) and can be a great asset to law enforcement. However, law enforcement agencies must take the time to ask themselves about the ethical considerations whenever using any kind of AI technology before jumping into this new field of research and development.
Recently, OpenAI released ChatGPT, a powerful autoregressive language model chatbot, and Google followed up with its own version. These interactive chatbots have the capability to quickly glimpse information from open sources, parse through context and generate human-like responses. Nevertheless, this ease of use, quick responses, and lack of supervision may lead to unethical use of technology.
When using AI chatbots to store data, it is important to remember to give it proper context in order to find better answers. AI chatbots respond more accurately and quickly if it is given the proper context to comprehend the question. If the context is manipulated, the chatbot can be tricked into finding even the most private information or lead to biased outcomes. As an example, when I asked ChatGPT to provide the personal information of my friend, it specified that it cannot access this kind of information without reliable sources.
Trainees and engineers should be aware of the potential biases in the system. AI chatbots learn from the data which has previously been fed into it, so the accuracy of the answers depend on the quality of the data. Furthermore, if the data is biased, the chatbot will be biased. For this reason, ChatGPT added a warning to inform users that it may generate incorrect or misleading information and produce offensive or biased content. Police should always remember when using any kind of AI tool to focus on criminal behavior rather than demographics, race, religion or sex.
In addition, transparency is crucial when using any AI technology for any police decisions. While there aren’t currently any laws that enforce the disclosure of this kind of erudition, each police agency should determine the best method to inform their citizens. Moreover, the chatbots aren’t fully secure and open-source chatbots may compromise user privacy and personal information. For that reason, users should not provide their personal identifying information to the chatbot to avoid any security issues.
AI has the potential to revolutionize the way police operate and complete investigations. It is essential for any police agency to recognize the ethical implications of any AI tool and its use before fully committing to it. Everyone involved should create guidelines and regulations to ensure that AI is used for the benefit of humanity, and not cause any harm. Having these conversations and debates is necessary to develop a technology that stands up to the highest standard of ethical use.
