As a part of TechCrunch’s ongoing Women in AI series, which seeks to give AI-focused women academics and others their well-deserved — and overdue — time in the spotlight, TechCrunch interviewed Dr. Rebecca Portnoff, who is vice president of data science at the nonprofit Thorn, which builds tech to protect children from sexual abuse. 

She attended Princeton University before receiving her PhD in computer science from the University of California, Berkeley. She has been working her way up the ladder at Thorn, where she has worked since 2016. She started as a volunteer research scientist and now, eight years later, leads a team that is probably one of the only in the world dedicated to building machine learning and artificial intelligence to stop, prevent, and defend children from sexual abuse. 

“During my senior year at Princeton, as I was contemplating what to do after graduation, my sister recommended I read ‘Half the Sky’ by Nicholas Kristof and Sheryl WuDunn, which introduced me to the topic of child sexual abuse,” she told TechCrunch, saying the book inspired her to study how to make a difference in this space. She went on to write her doctorate dissertation focusing especially on using machine learning and AI in this space. 

At Thorn, Portnoff’s team helps to identify victims, stop revictimization, and prevent the viral spread of sexual abuse material. She led the Thorn and All Tech Is Human’s joint Safety by Design initiative last year, which strives to prevent people from using generative AI to sexually harm children. 

“It was a tremendous lift, collaboratively defining principles and mitigations to prevent generative models from producing abuse material, make such material more reliably detected, and prevent the distribution of those models, services, and apps that are used to produce this abuse material, then aligning industry leaders to commit to those standards,” she recalled. She said she met many people dedicated to the cause, “but I’ve also got more gray hair than I did at the start of it all.” 

Using AI to create nonconsensual sexual images has become a big discussion, especially as AI porn generations become more sophisticated, as TechCrunch previously reported. There is currently no comprehensive federal law in place that protects or prevents sexual generative AI images created of other people without their consent, though individual states, like Florida, Louisiana, and New Mexico, have passed their own legislation to specifically target AI child abuse.

In fact, she said this is one of the most pressing issues facing AI as it evolves. “One in 10 minors report they knew of cases where their peers had generated nude imagery of other kids,” she said. 

“We don’t have to live in this reality and it’s unacceptable that we’ve allowed it to go to this point already.” She said there are mitigations, however, that can be put in place to prevent and reduce this misuse. Thorn, for example, is advocating that tech companies adopt their safety-by-design principles and mitigations, and publicly share how they are preventing the misuse of their generative AI technologies and products in furthering child sexual abuse, collaborating with professional organizations such as the Institute of Electrical and Electronics Engineers (IEEE) and the National Institute of Standards and Technology (NIST) to support setting standards for companies that can be used to audit progress, as well as engaging with policymakers to inform them of how important this is.

“Legislation grounded in impact will be necessary to bring all companies and stakeholders on board,” she said. 

As she rose through the ranks in building AI, Portnoff recalls people ignoring her advice, asking instead to speak with someone who has a technical background. “My response? ‘No worries, you are talking with someone with a technical background,’” she said. 

She said a few things have helped her navigate working in such a male-dominated field: being prepared, acting with confidence, and assuming good intentions. Being prepared helps her enter rooms with more confidence, while confidence allows her to navigate challenges with curiosity and boldness, “seeking first to understand and then to be understood,” she continued. 

“Assuming good intent helps me approach challenges with kindness rather than defensiveness,” she said. “If that good intent truly isn’t there, it’ll show eventually.” 

Her advice to women seeking to enter AI is to always believe in your ability and meaning. She said it’s easy to fall into the trap of letting the assumptions people have about you define your potential, but that everyone’s voice is going to be needed in this current AI revolution. 

“As ML/AI becomes more integrated into our human systems, all of us need to work together to ensure it’s done in a way that builds up our collective flourishing and prioritizes the most vulnerable among us.” 

Portnoff said there are many facets to responsible AI, including the need for transparency, fairness, reliability, and safety. “But all of them have one thing in common,” she continued. “Responsibly building ML/AI requires engaging with more stakeholders than just your fellow technologists.” 

This means more active listening and collaboration. “If you’re following a roadmap for building responsible AI, and you find that you haven’t talked to anyone outside your organization or your engineering team in the process, you’re probably headed in the wrong direction.” 

And, as investors continue to dump billions of dollars into AI startups, Portnoff suggested that investors can start looking at responsibility as early as the due diligence stage, looking at a company’s commitment to ethics before making an investment, and then requiring certain standards to be met. This can “prevent harm and enable positive growth.” 

“There is a lot of work that needs to be done,” she said, talking generally. “And you can be the one to make it happen.”