The future of artificial intelligence is often framed in spectacle: driverless cars, humanoid robots, machines that mimic human behavior. For Tuka Waddah Alhanai, the more meaningful shift is quieter and already underway.
“The shift isn't from human to machine,” she says. “It's from systems that wait for instructions to systems that understand intent.”
Alhanai, an Emirati engineer and assistant professor of computer engineering at NYU Abu Dhabi, is focused on what comes next: systems that don’t need to be told what to do.
Her work examines how machines can interpret what people are attempting in real-time, even in complex or partially visible environments. In practice, this could mean a system that recognizes a partially set table and completes the task, or notices someone struggling with a routine action and intervenes without being asked.
In most places, this kind of technology remains in pilot mode. In the UAE, it is edging closer to daily life, backed by investment in research, infrastructure, and policy.
The making of a machine learning pioneer
Raised in Abu Dhabi, Alhanai did not set out to work in artificial intelligence. “I was more generally interested in computers and how they work,” she says. “From very simple concepts like a circuit or a transistor, you can get a complex system like a computer.”
That curiosity led her to pursue her foundational training in electrical engineering in Abu Dhabi, and eventually to the Massachusetts Institute of Technology, where she completed her PhD and first encountered machine learning.
“I had never even heard the term machine learning before starting my program,” she says. “There were hundreds of researchers working on this, and I got really interested in the idea that you could take almost any problem and represent it as numbers.”
The possibility of translating human judgment into something measurable really drew Alhanai in. “There are some people who are really good at making certain kinds of judgments,” she says. “It could be clinical judgments, negotiating with another person, or predicting something about the stock market. I thought it would be amazing if you could use math to get close to that.”
Teaching machines to read the room
Today, Alhanai develops systems that model how people act and decide, particularly in situations where information is incomplete.
“One of the things we're interested in is when you have a human and a robot interacting,” she says. “The robot has to infer what the human is trying to do.”
Her team works with vision-language models, systems that combine visual input with language-based reasoning. “You can give it a video and ask what the action is, or what the objects are being assembled,” she says. “The goal is to understand what's happening in a granular way so it can decide the next best action.”
That shift, from reacting to anticipating, is what she sees as the defining change in AI. It moves systems closer to how people operate, reading context, filling in gaps, and making decisions without explicit instructions.
Where the world is headed
The implications extend well beyond robotics. “If you're a software developer, you're probably spending your whole day guiding AI agents,” she says. “The majority of the code is being created by the AI.”
She expects that model to spread, with systems handling processes while humans focus on judgment and oversight.
The UAE is well placed to accelerate that shift. “There's a lot of enthusiasm and positive sentiment towards AI innovation here,” she says. “At the industry level and at the academic level, there's a lot of activity and investment.”
Representing the UAE
Alongside her research and teaching, Alhanai was recently appointed to the newly established UN Independent International Scientific Panel on Artificial Intelligence. “It's a testament to the UAE's science and innovation ecosystem,” she says. “That there's a representative from the UAE is huge.”
The pace of change, she says, is relentless. “You go to bed thinking you have a sense of the AI landscape, and then you wake up and see something new has been released,” she says.
“It will be interesting to see how people and organizations adopt these tools, how much trust they have, and how people upskill.”
For now, the horizon is expanding quickly. “We're going to enter a scenario where you're much more exposed to interacting with robotic systems,” she says. “And the opportunities that come with that are enormous.”
