It’s time for investigators to adopt artificial intelligence. But it’s important to do it intelligently — with a mind to ethics, privacy, and transparency.
For professional investigators, AI is now a tool we should be adding to our toolbox — along with the search engines, calendars, invoicing, transcription, and investigative databases we already use. It’s no longer should investigators use AI, but investigators using AI should do so without compromising our skills, ethics, or evidentiary value.
Use It As a Productivity Resource (Not a Work Product Replacement)
AI supercharges tasks that can slow us down. It can rapidly read, summarize, and proofread reports, notes, and transcribed interviews. It can highlight patterns and inconsistencies in discovery materials — after the investigator has conducted their own review and analysis. It can extract bullet-point summaries from audio and video transcripts, including timestamps, which can be turned into chronologies, attorney synopses, and witness statements in minutes, not hours. These are legitimate, powerful uses.
But AI doesn’t have our evidentiary intuition, experience, and skills.
There are hard boundaries to be considered. AI should not be the primary fact-gathering or analysis source. A known risk is hallucination — i.e., when AI generates false positives, untrue “facts,” or conflicting answers to the same prompt.
Inconsistent parameters (prompts) can result in inconsistent conclusions. AI is also subject to human bias and GIGO (Garbage In, Garbage Out), making detailed prompt drafting as important as investigative discipline.
In fairness, we’re all subject to hallucinations and biases of our own and from many search sources — and we must verify everything. This should be second nature to investigators. And just as you wouldn’t turn in a printout of a database search result and call it your “report,” you shouldn’t paste a chatbot’s reply into a report and call it your analysis or conclusions.
AI replies may sound very convincing, but they’re no substitute for your investigative instincts and experience.
Use It As a Business Assistant: Get Help Analyzing Expenses, Price, and Value
New investigators too often use lower hourly rates to lure in business. The problem with this strategy is that low hourly rates don’t attract clients who actually value your expertise. You can never win in a race to the bottom on price. You’ll end up with clients who care more about low costs than actual value, and they’ll abandon you when they find a cheaper option.
With thoughtful prompting, AI can actually help you build your business model on the strength of value-driven rates. Ask your AI assistant to compare your current rate to suggested benchmarks based on your location, specialty, years of experience, and market conditions. Include information about your taxes, fixed overhead, project-specific costs, and desired profit margin. Then AI can help you calculate break-even points and realistic weekly billable hours and projections.
For example, you wouldn’t bill a two-day surveillance job simply by multiplying hours of surveillance plus mileage. You also have to account for case expenses, reporting time, video editing, and other applicable billable hours.
AI-assisted billing prompts will show flat-rate project costs exceeding the investigator’s simple estimates. They can help you nail down value-driven rates for retainers, identify cost savings opportunities, and calculate margins and markup to keep your business in the black.
Most investigators hate the business side of things. Let AI be your office assistant, leaving you more time to what you do best — investigate.
Best Practices & Ethical Considerations
If you plan to use generative AI as an investigative tool, transparency and privacy are paramount.
Transparency: Investigators, like attorneys, should always disclose whenever we use AI to assist in an investigation. And of course, we should always independently fact-check all AI-generated information. Treat AI as a secondary resource to refine and format your own work-product. As with any investigative resource, never trust, and always verify.
Privacy: With every client and case, investigators handle sensitive data. Maintaining confidentiality is essential to our clients’ legal cases and our professional reputations.
Be careful about what data you allow an AI platform to access. As with any other investigative resource and tool — including public records, databases, vendors, etc. — AI poses data privacy risks to users. The difference between AI platforms and those other resources is one of scale: as AI devours incalculable quantities of information, it’s hard to know what’s being gathered — with and without our consent — and used for training purposes; whether personal or sensitive information might be leaked to external sources; and how to control the collection and use of our data.
Investigators should look for AI platforms which are security enhanced and specific for investigative confidentiality. This article in Fast Company offers a quick overview of the default privacy settings of various AI assistants, including whether they use your data to train AI and how you can change settings to enhance security. (*note: This tech changes fast, so an article like this can fall out of date in mere months.)
Savvy Prompting: It’s Like Interviewing a Witness
Drafting AI prompts is about asking the right questions and giving detailed instructions. AI prompts should resemble a structured investigative interview. To get a specific type of response (such as a graphic, outline, research), give it very specific and detailed parameters when you ask it to perform a task. For example, here’s how I might ask it to create a comparison document from two or more database reports:
“Review the attached files and provide a bullet point synopsis of information I can follow-up with as a private investigator to contact [Subject] who is a defendant in civil litigation. I need the following: 1) Personal Identifying Information (PII) – Social Security Number (SSN), Date of Birth or Birthdate (DOB), and Last Known Address (LKA) with dates; 2) Primary or first-degree relatives of spouse, siblings, and children who live at the same address with their PII; 3) Employment with dates; 4) Current vehicles with owners, address, and dates; 5) Current property information with owners, address, and dates; 6) any liens, judgments, or bankruptcies.“
When investigators learn the art of prompting well, AI becomes more accurate, faster, and better formatted, and it increases its value as a resource. However, AI may misinterpret investigative and legal terminology; and all AI-drafted materials must be thoroughly fact-checked and reviewed for legal and ethical compliance.
Quick Studies of AI as a Resource
AI can also serve as a meta-search engine, combining its internal dataset with web searches: ask it to search its dataset, then expand with Google or other engines. There are benefits for potentially better accuracy, fewer tracking cookies, and no ad-cluttered results pages.
A test case: Let’s say you’re assessing the market for your services and rates. AI can perform market research to identify target markets, specialties, competing sites, and trends. It can also suggest improvements to website content, and SEO or tags for social media. Use it to test your agency’s SEO ranking with prompts like: “Find me a private investigator in [CITY, STATE]… Include Google and Yahoo… Exclude ads… Then expand the list.”
AI can also streamline tedious administrative tasks, including OCR and improved reading of PDFs, sometimes better than standard software. It can create templates for replies, agreements, workflows, and other documents. And from notes and transcripts, it can produce a bullet-point synopsis (third person), and first-person witness statements.
Of course, you should review all output thoroughly for errors and omissions.
Investigators: Adapt Intelligently
From typewriters to fax machines, from multi-function printers to cloud storage, technology evolves and we adapt. AI is now the Next Big Thing, and it’s here to stay. Investigators either harness it or risk losing to (or being undercut by) AI-driven services and competitors.
Our value isn’t diminished by the emergence of AI as a resource — IF we adapt intelligently. AI is a tool. Its output should never become our work product, any more than the results of a database search should.
AI is here to stay, as are the investigators who adapt to it intelligently — i.e., with a mind to ethics, privacy, and transparency.
About the authors:
Dean A. Beers, CLI, CCDI founded Associates in Forensic Investigations, LLC in 1987. Karen S. Beers, BSW, CCDI joined the agency in 1996. Since 2008, the two have specialized in expert consultations, legal investigations, personal injury, and negligence & death in civil, criminal and probate litigation. Together they developed the Certified Forensic Death Investigator (CFDI) Program exclusively for criminal defense investigators.
Dean and Karen are members of several associations, including Professional Private Investigators Association of Colorado, National Association of Legal Investigators, Criminal Defense Investigation Training Council, World Association of Detectives, and others.
Want more detail? See Dean & Karen Beers’s new course: AI for Private Investigators, created for new and veteran private investigators looking to learn the basics and improve their use of AI for investigative administration and cases. You’ll learn outside-the-box uses for AI in marketing, admin tasks, and work product.
Improve your time management and develop savvy query prompts to support your investigative strategies, use your resources better, and earn your value.