Institute of Artificial Intelligence of China Telecom. Rooted in AI, flowing across edges.
Joined April 2026
- Tweets98
- Following35
- Followers126
- Likes0
👂Heard @Apple’s cooking up a "no-recording" camera for privacy?
It's a solid direction, no doubt.
🤷Just a few months late to the party.
🔉We already built it. Nearly A YEAR AGO!
🚧For years we've been stuck in an impossible tradeoff:
🚬Smoking, bullying, or accidentally falling often take place in high-privacy spaces— but filming people there is a hard no.
🔧Our solution: AI spots risky behavior without exposing who's involved.
Under the leadership of Professor Dr. Xuelong Li, CTO and Chief Scientist of China Telecom and Director of TeleAI, TeleAI Privacy Camera has now advanced to a new version following its unveiling in February 2026, - and is fully integrated into the TeleCare (TeleAI's Smart Care Assistant).
#PrivacyByDesign #ResponsibleAI #AIFlow #TeleAI #J450 #apple
📷TeleAI Privacy Camera turns visual information into semantic or comic-style illustrations, enabling risk alerts without exposing identifiable information.
🙈Behavior becomes visible. Identity stays private.
🚬 Someone smoking in a restroom? Their posture and actions are clear, without showing their real appearance.
🚨 Someone falls on campus? The illustration helps staff quickly grasp what happened.
🏥 A resident needs help in a care home? Staff can better understand the situation while limiting exposure to private visual details.
It works anywhere safety matters but privacy is non-negotiable.
🕶️Less exposure. Clearer understanding. Privacy by design.
A camera doesn't need to reveal every detail to tell us what happened.
📡Powered by AI Flow, TeleAI's Privacy Camera turns visual signals into actionable insights — so you can act on risks fast, while personal details stay completely out of sight.
📌See the behavior. Understand the situation. Respect the privacy.
That's a smarter way to think about AI-powered safety. And it's only the beginning...
Teaching Robots to “Dream” Up Future 3D Scenes: Meet WNM-3D
Every move we make starts with imagination.
Before you turn a corner, what’s waiting on the other side?
🚧A dead end?
🛣️A new path?
🚴Or a cyclist rushing straight toward you?
🤖Robots need to think ahead, too.
Not just predict the next action,
but imagine the world that action will reveal.
Led by Professor Xuelong Li, CTO and Chief Scientist of China Telecom and Director of TeleAI, the TeleAI research team introduces an innovative solution:
WNM-3D
A framework that lets embodied AI look ahead before it acts.
#EmbodiedAI #Robotics
WNM-3D can imagine several steps ahead.
But it doesn’t commit to the whole plan.
Instead, it uses receding-horizon closed-loop inference:
🔁Observe → Model → Imagine → Plan → Act → Observe again
The model predicts a short stretch of the future, but only executes the next few actions.
Then reality takes over.
🔑New view? → Update the 3D scene memory.
🔑New state? → Replan.
Think ahead. Act step by step.
📈On GN-Bench, WNM-3D shows strong closed-loop performance across both seen and unseen environments, using only monocular RGB as visual input.
But what if this world-model intelligence could go beyond a single robot?
🌐That’s where AI Flow comes in.
Device → Edge → Cloud
📱The device handles real-time perception and action.
⚡️The edge provides fast, nearby reasoning and coordination.
☁️The cloud brings deeper world modeling, learning, and knowledge.
🤖One robot can imagine the world.
AI Flow lets that intelligence flow beyond the robot.
#WNM3D #AIFlow #Robotics
✨What makes a robot truly intelligent?
Not just knowing what to do,
but grasping what happens next.
Traditional navigation asks:
“Given what I see, where should I go?”
✅WNM-3D goes one step further:
“If I go there, what will I see?”
And with AI Flow, this predictive foresight is not confined to a single robot—
🤝It can flow, collaborate, and continuously improve across the physical world.
#AIFlow #EmbodiedAI #Robotics
Beneath the Surface, Beyond the Limits: Wireless Underwater Video Transmission
🌍71% of our planet lies beneath the waves.
Yet communication barely scratches the surface.
🌊Beneath the waves, communication faces fundamental barriers:
- Brutal signal attenuation
- Extremely limited bandwidth
These constraints have long limited nearly every form of underwater transmission.
For decades, even images struggled to travel underwater.
🎥Now, high-definition video can flow through the water in real time.
Under the leadership of Prof. Dr. Xuelong Li, CTO and Chief Scientist of China Telecom and Director of TeleAI, the TeleAI research team is pushing the boundaries of underwater connectivity with AI Flow.
From unreachable depths to connected intelligence.
#AIFlow #TeleAI #UnderwaterCommunication #AICommunication
With the acoustic bandwidth bottleneck broken,
our self-developed cross-domain air-sea drone now delivers:
✅ Real-time 720p@24fps HD video
✅ Wireless communication beyond 10 meters underwater
✅ Air-to-sea cross-domain link
This opens new possibilities for underwater exploration, ecological monitoring, and emergency search and rescue.
#AIFlow #UnderwaterAI
Space, air, land, sea — the loop is now complete.
🛰️Space unlocks the unreachable.
🚁Air expands the horizon.
🤖Land enables intelligent collaboration.
🌊Sea reveals the unseen.
Breaking boundaries between environments, AI Flow is building the communication foundation for the AI era:
not just connecting devices,
but connecting intelligence.
#AIFlow #AICommunication #FutureNetworks
What If You Could Control Robots Thousands of Miles Away?
🤖A single robot can solve one problem.
But when robots connect, the possibilities are boundless.
We've explored intelligence soaring in space and in the air.🛰️
Now, let's bring it down to earth.
Here, embodied intelligence takes center stage.
To move from the lab into the real world, two things matter most:
👤↔️🤖
Human–robot communication for remote operation
🤖↔️🤖
Robot–robot communication for multi-agent collaboration
Both hinge on one thing: intelligent communication.
Guided by the vision of Professor Xuelong Li, CTO and Chief Scientist of China Telecom and Director of TeleAI, the TeleAI research team is bringing AI Flow to embodied intelligence
#AIFlow #TeleAI #EmbodiedAI #Robot
🤔What if multiple robots could work together on a complex task—without a human in the loop?
With AI Flow’s unified token-based communication, robots share situational awareness and task knowledge seamlessly.
Think of it as a group chat for robots — but one built for real-time action.
With minimal bandwidth cost, robots collaborate more efficiently — unlocking capabilities like:
✅ Skill sharing
✅ Sim-to-real transfer
✅ Self-learning
From Bit Flow to Token Flow.
From connected robots to coordinated intelligence.
#MultiAgentSystems #RobotCollaboration
For embodied intelligence, connection is only the beginning.
🔁From human–robot remote operation to robot–robot collaboration, AI Flow is connecting intelligence across physical boundaries.
Now the flow has traveled across space, air, and land.
🌊It is ready to converge with the sea.
🌐Join us for the final chapter,
as AI Flow completes its journey toward a fully connected intelligent world.
#AICommunication #AIApplication
Less Bits, Sharper Visuals: The Drone-satellite Transmission over a Voice-call Bandwidth
⚡️A drone can be light. Its data shouldn't have to be.
✈️As part of AI Flow’s journey across space, air, land, and sea, we’re turning to the air.
The smaller the drone, the tighter the limits on payload, power, and connectivity.
🤔So how can a lightweight platform still transmit rich information over a narrow link?
Under the guidance of Prof. Dr. Xuelong Li, CTO and Chief Scientist of China Telecom, Director of TeleAI, the research team of TeleAI offers a new answer.
#AIFlow #TeleAI #GenerativeTransmission
📊Fieldtested performance:
- AI Flow Generative Transmission: average 60 Kbps
⚔️VS
- Conventional H264 encoding: average 1.5 Mbps
The result?
Real-time HD video streaming from places with weak network.
#AIFlow #DroneTech #SatelliteCommunications
Where can this capability make a difference?
- 🚒Natural Disaster response
- ⛰️Remote areas
- 🌊Oceanwide communications
In these network-constrained environments, AI Flow helps small drones fly farther, see clearly, and transmit reliably.
This is more than drone video transmission.
It’s another step toward AI-native communication across space, air, land, and sea.
🔔Please keep pace with our cross-domain exploration.