A team at Tufts University developed NeuroBridge to teach non-autistic people how autistic communication can differ. Autistic people often rely less on body language and may interpret sarcasm or figurative phrases literally, so they may prefer direct, clear language.
NeuroBridge uses large language models to create conversational scenarios and offers three response options at key moments. The options have similar meaning but differ in tone, clarity or phrasing. In one example, two options invite a yes/no answer while the third clearly asks for advice.
The researchers, including PhD candidate Rukhshan Haroon and associate professor Fahad Dogar, said the tool is not an on-demand translator. They worked with a board of autistic volunteers and tested the system with participants, who reported that feedback made it easier to see how parts of a conversation could be received differently. The team plans to explore campus support and further evaluation.
Difficult words
- autistic — a person with a neurodevelopmental condition affecting communication
- body language — nonverbal signals such as gestures and posture
- sarcasm — saying the opposite of what is meant
- figurative — language that uses non-literal meanings
- literally — in the exact, usual or actual meaning
- large language model — a computer system that generates human-like textlarge language models
- feedback — information given about performance or actions
- tone — the speaker's attitude or style of speaking
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- How might NeuroBridge help students and staff on a university campus?
- Have you ever misunderstood someone's meaning because of tone or phrasing? What happened?
- Do you prefer direct or indirect language when someone asks for help? Why?
Related articles
AI expands sexual and reproductive health access in Latin America
Research groups in Peru and Argentina use AI tools to give sexual and reproductive health information to young and marginalised people. Experts praise potential but warn of bias and call for better data, rules and oversight.
Touchscreens on car dashboards increase driver distraction
A simulator study found that using a car touchscreen while driving makes steering and touchscreen tasks worse. Multitasking reduced lane control and touchscreen accuracy; researchers suggest simple sensors could monitor attention and change the interface.
AI and Wearable Devices for Type 2 Diabetes
A meta-review from the University at Buffalo examines AI-enhanced wearable devices for Type 2 diabetes and prediabetes. The study finds predictive benefits and important limits, and calls for larger, more transparent studies before routine clinical use.