Lindsay Lieberman. (Photograph by Sandra Morlet)
Just a few months in the past, I attended a authorized convention the place synthetic intelligence was a dominant matter of dialogue. As an legal professional who represents victims of intimate picture abuse and gender-based violence, I used to be keen to listen to how the career was serious about AI.
One packed presentation within the convention’s largest ballroom centered on AI-generated proof. 4 presenters walked attorneys by way of three eventualities during which girls used AI to manufacture proof towards males. The eventualities have been illustrated with AI-generated movies depicting the ladies assembling their false instances, set to ominous, crime present music. One concerned a office sexual harassment declare; one other concerned a custody dispute. At one level, a presenter described the lady within the custody situation as utilizing AI to “destroy a person’s life.”
Because the presentation went on, I grew more and more uncomfortable with its repeated concentrate on girls fabricating claims. Over the previous 15 years, I’ve labored with survivors navigating office harassment and abuse in household, civil and legal issues. I’ve seen how exhausting they have to battle to have their proof taken critically. I’ve by no means encountered a survivor who used AI to fabricate it.
After the presentation, I approached one of many panelists and requested how usually individuals have been really utilizing AI to manufacture proof in instances like these. He couldn’t level to at least one documented case. The eventualities, he defined, have been meant for example what AI may do.
Lots of of attorneys had spent one hour and 20 minutes contemplating eventualities during which girls fabricated proof of abuse. I’d later current on the hurt AI is inflicting girls and ladies—however to a smaller viewers in a a lot smaller room.
In one in all my instances involving nonconsensual intimate picture dissemination, legislation enforcement initially informed my consumer they might not make an arrest with out an IP tackle linking the perpetrator to the emails, regardless that sturdy circumstantial proof recognized him because the sender. I needed to persuade investigators to look at that proof fairly than deal with the lacking IP tackle as the top of the inquiry. When investigators already overlook out there proof as a result of one digital identifier is lacking, the unsupported suggestion that proof “might be AI” dangers turning into one other excuse to cease investigating.
Digital proof is already troublesome for victims to elucidate and for the authorized system to guage. An unsupported suggestion that proof “might be AI” might give decision-makers another reason to cease wanting intently.
In the meantime, persons are utilizing AI instruments to create sexualized photographs of my purchasers with out consent. These photographs are used to humiliate, threaten and extort. Victims are left explaining to legislation enforcement that graphic photographs are pretend whereas coping with the very actual injury brought on by their circulation. The pictures are pretend. The abuse will not be. The proof is actual.
These are the issues legal professionals must know the right way to tackle now. These usually are not hypotheticals.
The subjects we give a platform at authorized conferences form how attorneys argue instances and the way judges consider them. If legal professionals depart with the impression that fabricated proof is a typical characteristic of abuse instances, protection attorneys might elevate unsupported claims that genuine proof was made with AI, and judges or jurors might give these claims extra weight than they deserve. A survivor may then be compelled to defend the authenticity of actual proof earlier than anybody addresses the abuse it paperwork.
That concern is particularly acute in sexual harassment and abuse instances. Survivors already face questions on their credibility, their recollections and why they got here ahead after they did. They need to not face a brand new presumption of doubt just because AI may have been used to create proof.
Attorneys, legislation enforcement and judges want to grasp the right way to consider digital proof: the place a file got here from, what its metadata can and can’t inform us, and the right way to think about different proof alongside it. That features circumstantial proof of who was accountable, equivalent to motive, alternative and conduct earlier than or after the incident.
There may be huge demand for AI programming, and authorized convention coordinators are understandably keen to satisfy it. However the urgency to show about AI makes it extra essential to scrutinize the examples we select and the assumptions these examples depart with an viewers.
Authorized conferences have a accountability to vet their audio system, ask whether or not the eventualities they current replicate documented issues, and provides actual harms the eye they deserve. Attorneys want AI training grounded in proof and context, with out presenting gendered hypotheticals that fail to replicate actuality.
Lindsay Lieberman is the founding father of Lindsay Lieberman LLC, a legislation and consulting company, the place she represents victims of intimate picture abuse and gender-based violence. A former Brooklyn, New York Metropolis, prosecutor, she has spent 15 years working with survivors throughout legal and civil issues.
ABAJournal.com is accepting queries for authentic, considerate, nonpromotional articles and commentary by unpaid contributors to run within the Your Voice part. Particulars and submission pointers are posted at “Your Submissions, Your Voice.”
This column displays the opinions of the writer and never essentially the views of the ABA Journal—or the American Bar Affiliation.