AI tools have gained prominence in recent years, but as a self-proclaimed AI sceptic, I have been slow to explore their capabilities. I trust my own skills and believe that authenticity adds value—especially when providing professional opinions or applying for jobs. However, despite my reluctance, I recently found myself relying on AI during my ongoing custody dispute.
The Role of AI in a Custody Battle
My journey with AI began when I discovered that Skype was shutting down. While I had protested its use for video chats with my daughter, our court order required all communications to be conducted through Skype. So when Microsoft announced it would deprecate Skype on May 5, 2025, I was caught off guard.
Knowing that all records of my communications with my ex were stored on Skype, I quickly searched for a way to download them. This led me to a large JSON file containing over ten years of detailed records. Around this time, my lawyer asked me to recall historical details, requesting exact dates and specific evidence where possible. Given the sheer volume of breaches since the court order was established, it was difficult to keep track of everything accurately.
Determined to find a solution, I explored tools that could analyse JSON files. After struggling to find the right fit, I decided to use Copilot at work, typing in basic instructions and seeing what it could do.
My Initial AI Experience
To my surprise, the first AI-generated document Copilot produced was fairly accurate. While the formatting had issues—such as ugly code-style blobs—it was impressive overall. AI didn’t magically tidy things up when I asked, but with some practice, I managed to organize the data into a clear, structured format without needing spreadsheets or pivot tables.
Human oversight was still necessary. For example, exporting calls and messages separately was more efficient, as combining them led to confusion. Additionally, I needed to review the final document carefully, as some of the AI-generated assumptions were inaccurate. Despite these limitations, the process felt intuitive.
Another AI-powered task I attempted was translating poorly scanned PDF documents. Optical character recognition and online translation tools have been fairly reliable for years, but Copilot simplified the process by providing a one-stop solution. Previously, I would have needed multiple applications to achieve similar results.
Creating an AI-Assisted Court Submission
After using AI to analyse and structure data, I had a solid foundation for my court statement. Without sharing too many sensitive details, the document outlined thousands of breaches that had occurred since the court order was issued. While I had always suspected noncompliance, this data quantified every instance in stark detail. Given the overwhelming amount of information, I needed a way to summarize everything effectively.
Previously, I would have relied on pivot tables and charts to create ‘stories’ from data. This time, I asked Copilot to generate a summary in the style I wanted. The result condensed hundreds of pages of chat logs and phone records into clear statistics—showing the number of calls I made, how many were ignored, how many were successful, and how often my daughter’s stepfather refused to let me speak with her. It also validated that my calls were compliant with the court order, adding credibility to my case.
With this summary in hand, I could confidently craft my own narratives. While I could have recalled events from memory, backing my claims with concrete data—specific dates, times, and direct quotes—was empowering.
The Legal Utility of AI
It remains uncertain whether my AI-assisted narratives will persuade the court. However, experience has taught me that opposing parties often challenge arguments by disputing minor factual details—such as exact dates or whether a conversation took place. The ability to provide precise statistics on call refusals and common excuses, while maintaining full confidence in the underlying data, offers significant advantages.
Final Thoughts on AI
This experience has made me appreciate AI’s capabilities in data formatting and analysis. Beyond the hype surrounding ‘man versus machine’ debates (which I believe are often exaggerated for storytelling purposes), I valued the convenience of automating tasks without needing to purchase separate software.
One of the most frustrating aspects of preparing evidence for child abduction cases is the emotional toll of analysing years of conflict and trauma. It is time-consuming and mentally exhausting—particularly for those who have developed mental health conditions from the stress. AI alleviates some of this burden by simplifying workflows.
Additionally, dealing with outdated technology and multiple languages adds complexity to legal documentation. Previously, I would spend entire weekends manually reviewing records, sometimes breaking down from the emotional weight of it all. AI doesn’t eliminate this burden entirely, but it provides structure and efficiency that can help ease the process.
That said, AI is not ‘intelligent’ in the way many imagine. It cannot take a pile of documents, interpret a lawyer’s instructions, and generate a polished court submission without guidance. AI still requires constant human direction. However, it represents a quantum leap from traditional workflows, removing much of the technical complexity involved in data analysis. For instance, processing JSON files and structuring bilingual legal records—tasks that once required advanced expertise—can now be accomplished within minutes.
I am relieved that future left-behind parents will have AI tools at their disposal. While AI is merely a tool, it empowers us and our children to present structured, factual accounts when fighting for custody. It is not a silver bullet, but in a world where abductors use gaslighting and selective memory to gain an advantage, AI provides crucial utility.
