My inbox is full of messages from experts cold emailing me asking to be a source for articles I’m working on. So the message from Sophie Kress, inviting me to interview her for a future article, seemed typical at first. “I have over 8 years of experience and qualifications in the field of psychology and couple & family therapy and am passionate about exploring possibilities for collaboration, especially in the area of love, relationships or LGBTQIA+ topics. ,” Kress wrote. She included a list of links to articles where she was featured as an expert. But I stopped because her email address used the domain name SexalAlpha.com, a website that reviews sex toys.
Cress’ therapy website described her as having a master’s degree in couples and family therapy from Antioch University and being a licensed marriage and family therapist in North Carolina. The site also lists two other certifications (Prepare/Enrich and Gottman Therapist) that many marriage and family therapists have but are less known to the general public.
But to me, as a freelance reporter who frequently writes about mental health and has been working in therapy for 20 years, the tone seemed different from other therapists’ websites. The biography on this site begins with “Hello, my love, it’s so wonderful to have you here!” The most prominent headshot on the site shows a woman in a superhero pose, looking upwards and away from the camera, a sly smile on her lips.
I came to this website to learn more about Cress credentials, but instead found more questions.
Since generative AI programs like ChatGPT entered the public consciousness in late 2022, many of our digital experiences have taken on a new level of mistrust. Is it really good advice to put a little glue in your pizza sauce to prevent the cheese from sliding off? Is Billy Corgan actually promoting the new Smashing Pumpkins album in eight languages? Really you? ‘s mother called to say she was being held captive and that the kidnappers would kill her if she didn’t pay the ransom?
For journalists, already professionally required to be skeptical, that distrust must extend to potential sources as well. Of course, with all due respect to prostitutes, grifting may actually be the world’s oldest profession. But with AI programs, it’s incredibly easy to generate chunks of text that look like they were written by experts in every conceivable field, at least on first skim.
At the same time, telephone calls began to take the form of answering machines (remember before…). In our personal lives, text is faster. In our work, many people use email to communicate with external parties. And if you’re a multi-tasking reporter, email interviews may be appealing. It’s often faster and more convenient than scheduling a phone call for both journalists and sources. A site that some reporters (including this one) use to facilitate this type of interview is Help a Reporter Out (HARO). The site connects journalists with self-proclaimed experts who can comment quickly, often by email. When I first heard from Cress, she emailed me directly and said she had “seen[my]website” linking to my writing portfolio. But the second time, she responded to one of my HARO requests for a story I was writing for another medium.
(While I was reporting this article, HARO became Connectively, a platform owned by PR-focused software company Cision. Last December, Cision shut down Connectively, but a spokesperson According to CisionOne, the decision was made to allow the company to focus on () CisionOne, which provides media monitoring and “journalist support tools” for marketing and PR professionals. The company’s platform Unless otherwise noted, all of the company’s responses to my requests for comment were made before Connectively was shut down. )
I responded to Ms. Cress’ email pitch and told her I wanted to interview her for an article about how to find a legitimate mental health professional online. A few days later, she responded and agreed to an interview, but was told it could only be conducted via email, even after I requested a phone or video conference. “Sorry, I’m always on the go, so we can’t chat by phone or Zoom at this time, so email is the best option for now,” she explained. When I told her that I really, really could only do the interview via Zoom or phone, she stopped responding.
Of course, one of the problems with email interviews is that it leaves open the possibility that the sender is not the person writing the response. The rise of generative AI chatbots leaves open the possibility that the replies may not have come from a human mind. At least one writer posting on Reddit lamented that some of the answers he received from Connective Source experts appeared to be written by an AI. Of course, just as poorly written writing doesn’t necessarily mean it was created by a robot, the use of AI also means that the source of the information doesn’t accurately represent its credentials. It doesn’t inherently mean that. People from a variety of disciplines report integrating generative AI tools into their work. There are also many reports of people using AI to spread misinformation, revenge porn, and really bad art. As is often the case with emerging technologies, it remains to be seen whether the benefits outweigh the drawbacks.
