"The most empathetic voice in my life"

For Kate D'hotman, a Cape Town-based horror film director, connecting with film audiences comes naturally. For her, talking to people is much more terrifying. "I've never understood how people decipher social cues," says the 40-year-old.

D'hotman has autism and attention deficit hyperactivity disorder (ADHD), which can make communicating with others exhausting and challenging. But since 2022, she has been regularly using ChatGPT, a popular artificial intelligence chatbot developed by OpenAI, to overcome communication barriers, both at work and in her personal life. "I know it's a machine," she says. "But sometimes, honestly, it's the most empathetic voice in my life."

Neurodivergent individuals, including those with autism, ADHD, dyslexia, and other conditions, experience the world differently than the neurotypical norm. Talking to a colleague, or even texting a friend, can lead to misinterpreted signals, misunderstood tone, and unwanted impressions.

AI-powered chatbots have become unexpected allies, helping people navigate social situations with real-time advice. While the new technology carries certain risks, especially when it comes to potential overreliance, many neurodivergent users now see it as a lifeline.

What does this look like in practice? For D'hotman, ChatGPT is an editor, a translator, and someone she confides in. Before she started using the technology, she says that communicating in "neurotypical" environments was extremely difficult for her. She remembers once, at the request of her superior, she sent a list of items to improve the company's business. However, what was for her a simple and direct message was received as too harsh, even rude.

She now regularly uses ChatGPT to check the content of her messages, asking the chatbot to take into account the tone and context of the conversation. Sometimes she casts it as a psychologist or therapist, asking for help navigating sensitive situations, such as a disagreement with her best friend. Once, she even uploaded months of correspondence between them, asking the chatbot to help her spot things she might have missed on her own. Unlike humans, D'hotman says, the chatbot is always positive and non-judgmental.

While artificial intelligence can improve efficiency, it can also reduce critical engagement, especially in routine or lower-risk tasks, where users simply rely on AI, researchers say.

It’s a feeling that other neurodivergent people recognize. Sarah Rickwood, a senior project manager in the sales training industry from Kent, England, has ADHD and autism. She says her mind often “runs off,” and people in conversations can’t follow her. “I don’t present myself properly,” she says, noting that ChatGPT has allowed her to “do so much more with my brain.” With its help, she can compose emails and business proposals more clearly.

The use of AI-powered tools is skyrocketing. A study conducted in January by Google and research firm Ipsos found that worldwide use of AI had jumped 48%, with enthusiasm for its practical benefits now outweighing concerns about its potential negative effects. In February, OpenAI told Reuters that its weekly active users had surpassed 400 million, at least 2 million of whom are business subscribers.

But for neurodivergent users, these aren't just convenience tools - certain AI-based chatbots are now being developed with this very community in mind.

Michael Daniel, an engineer and entrepreneur from Newcastle, Australia, told Reuters he didn't realize how long he had been hiding his neurodivergent traits until his daughter was diagnosed with autism - and then he was. The desire to express himself more clearly in communication with his neurotypical wife and loved ones inspired him to develop Neurotranslator, an artificial intelligence-based personal assistant that he credits with helping him fully understand and process social interactions, as well as avoid misunderstandings.

"Wow... that's a really unique shirt," he recalls once saying of his wife's wardrobe, not realizing how such a comment could be interpreted. His wife asked him to run the sentence through the Neurotranslator, and it was then that he realized that, in the absence of positive affirmation, comments about someone's appearance can come across as criticism.

"The emotional burden of those situations would just disappear in a matter of minutes," Danijel says of using the app. Since NeuroTranslator launched in September, Danijel says it has gained more than 200 subscribers. The app's previous web version, called Autistic Translator, had 500 monthly subscribers.

While the technology is already incredibly transformative, some warn of the dangers of relying too much on it. The prospect of having an answer at any moment can be "very seductive," says Larisa Suzuki, a London-based computer scientist and NASA visiting researcher who is neurodivergent herself.

Too much reliance could be harmful if it interferes with the ability of neurodivergent people to function without such a tool, or if the technology itself becomes unreliable, as is already the case with many AI-powered search results, according to a recent study in the Columbia Journalism Review. “If AI starts making mistakes and providing incorrect information,” Suzuki says, “people could lose faith in the technology — and in themselves.”

Opening your soul to an AI chatbot certainly carries risks, agrees Gianluca Mauro, an artificial intelligence expert and co-author of the book Zero to AI. “The goal [of a model like ChatGPT] is to satisfy the user,” he says, questioning the AI’s willingness to offer critical advice. Unlike therapists, these tools are not bound by ethical codes or professional standards. If AI has the potential to be addictive, Mauro adds, regulation should be considered.

A recent study by Carnegie Mellon University and Microsoft (a key investor in OpenAI) suggests that long-term overreliance on generative AI-based tools can undermine users’ critical thinking skills and leave them unprepared to function without them. “While AI can improve efficiency,” the researchers wrote, “it can also reduce critical engagement, especially for routine or low-risk tasks where users simply rely on AI.”

While Dr. Melani Kacman, a clinical psychologist and human behavior expert, recognizes the benefits of AI for neurodivergent individuals, she also points to potential negative consequences - such as it can serve as an excuse for patients to avoid contact with others.

The therapist, she says, encourages the patient to try new things outside their comfort zone. “I think it’s much harder for your AI companion to push you to do that,” she says. But for users who have already relied on the technology, such warnings feel academic and distant.

"Many of us end up just withdrawing from society," warns D'hotman, who says she barely left the house for a year after her autism diagnosis because she felt overwhelmed. If she stopped using ChatGPT, she fears she would return to that traumatic period of isolation. "As someone who has struggled with the disorder her whole life," she says, "this is essential for me."

Translation: NB