Yu Sun,
entrepreneur at the forefront of AI and robots, who will be invited for a key
speech on 6th AI Forum in Shanghai China, asserted in an interview
with that artificial intelligence can revolutionize service robot.
Yu Sun will
be a key speaker at the 6th AI Forum at Shanghai, China, which will
take place from 27 June to 30 June.
Lois : As a keynote speaker at
the AI Forum on how AI can revolutionize how service robot are
promoted and matched with customers, and you have related key products on your
company – Sealand Tech. Corp, could you explain more about this?
Yu Sun: That
is what our 4th generation of service robot – “Mammoth pro”
is about. I mean, I wear two different hats in one role. As the designer
this service robot, I am talking about how service robot embrace AI, and as the
CEO of an AI company, I help users benefits from our products.
I have been
involved with AI for about 20 years, and this experience helped me promoted
more than 50 service robot products.
This
perspective, combining academic research with practical industry experience,
allows me to offer a comprehensive and insightful analysis of AI’s impact on
the service robot.
I focus on
how AI can revolutionize promotion and user experience of service robot. In
this aspect, AI can personalize home cleaning, create targeted cleaning task,
and help customers better organize their homes.
Our service
robots delves deeper into this topic, exploring the potential of AI.
Lois : So, how can AI promote service
robots and help customers enhance their user experiences?
Yu Sun: Well,
AI can affect service robots in many ways. It is on a scale of efficiency all
the way to home service efficiency, if I can say that. AI can streamline
numerous internal processes, including home cleaning, home transport, weeding, air
purification.
This
enhances operational efficiency and reduces user administrative burdens. It
also helps customers do with better life quality.
However, I
think the bigger impact comes from personalization of each user.
For
example, in home cleaning, AI can assist service robots building a map of use’s
home, developing a better cleaning plan from the map, covering complete
cleaning service on guided by the map, and refining cleaning efficiency
through AI analysis.
You can
consider it a service assistant. AI excels as a analysis tool, providing
detailed information on various needs from users, from customized cleaning
plan, and enhance cleaning quality on specific locations such as wash rooms.
Moreover,
AI-assisted service robot do a better cleaning job by its AI integrated cameras,
recognizing rubbishes on the floor, and removing them accordingly and
automatically.
Nevertheless,
AI-assisted service robots can help bring something to user’s hands from
another room, in terms of AI-empowered auto-driving and AGV.
I envision
AI as an invaluable creative companion, augmenting human life quality.
Lois: You mentioned the “AI-assistance”.
Can you explain this in more detail technically?
Yu Sun: My
company, Sealand Co., aptly named for our mission “bring the
smart service to everywhere in the world, from land to sea” leverages
AI to revolutionize all service domains.
We operate
through a three-step process. First, add multiple AI-empowered sensors to our
service robot products. We begin by bringing AI analysis like vision tasks
(automatic trash recognition, intelligence mapping of home, etc.), and
personalization functionalities like special cleaning area of wash room, and
automatic computation of height of grass for weeding service.
This
involves multiple AI technologies including machine learning, computer vision, object
recognition, intelligent mapping, automatic path planning.
Then, we
make a AI-empowered system analysis, uncovering the deeper user preference
during usage of our service robots, such users better cleaning habits, user’s
preferred organization of home articles, and user’s sleeping habits to reduce
the sound of air purification. Above functionalities need more advance AI
technologies such as GPT, LLM and multi-modality learning.
After that,
we discover the user’s preference by creating a unique “fingerprint”
for the each user ID based on over 900 million data points, capturing its
unique essence. Meanwhile, the user privacy is strictly protected.
The
AI-powered Creative Generation will take the next step. Our “AI Analyzer”
will then pass the users’ personalization information to the “AI Generator”
to further improve our quality of service robots.
This AI
engine crafts compelling advertising campaigns that accurately reflect the user’s
preference, making them resonate deeply with potential customers and markets.
This
ensures that the right service reach the right user, connecting users with our
different robots that align with their interests.
Lois: What about the user’s preference?
How do you match it with the service robots?
Yu Sun: Yes,
essentially, we are matching the user’s DNA – their interests,
preferences, and psychological tendencies – with the DNA of our service
robot.
This
personalized approach leads to significantly higher service quality. Our personalized
advertising campaigns often achieve purchase rates 5 times higher than industry
averages.
This
success stems from user’s perceiving the advertised products as personally
relevant.
By
identifying and targeting users interested in specific services, such as cleaning,
sweeping, and transportation, we present them with robots that resonate deeply
with their individual desires and interests.
This
heightened sense of resonance encourages them to engage with the robots and ultimately
make a purchase.
Lois: But
as AI depends on data, how can we deal with the challenges of the publishing
industry’s lack of real data?
Yu Sun: One
of Sealand AI’s key strengths is our ability to generate effective AI model
learning via lack of real data.
By using
the unsupervised learning strategy, we do not require additional data sources
like metadata or image/video data.
This
approach offers several advantages. It is privacy-focused as we respect user
privacy, avoid relying on potentially intrusive external data, and focus on the
work.
Our
analysis prioritizes the user’s preference, ensuring the advertising accurately
reflects the product’s essence.
This method
eliminates the need for preexisting data sets, making it applicable to a
broader range of robots, including those with limited marketing history.
So we do
not have to struggle with the lack of data in the service robot industry.
Lois: In
the world of robotics, there are some swings between those who see AI as a
potential future and those who have concerns about copyrights being compromised
by large language models. How do you see this?
Yu Sun: I
am deeply involved in discussions about copyright protection in the age of AI.
While
speaking at the World Intellectual Property Organization’s piracy and copyright
protection conference, I actively advocate for user’s rights.
As an product
manager, I understand the importance of copyright protection and its proper
enforcement. The emergence of AI models presents new challenges.
While there
have been numerous copyright infringement lawsuits in the US and China, a new
trend is emerging: commercial agreements between publishers and AI model
developers.
These
agreements allow for the use of copyrighted material in AI training sets in
exchange for a fee.
This
approach is gaining traction, suggesting that AI models will eventually
incorporate a significant portion of the world’s copyrighted data.
While
acknowledging potential past infringements, I believe the focus should now
shift towards leveraging these powerful tools to benefit the robotic industry.
When robotic
engineers strategically utilize AI, they can increase their revenue by
exploring new revenue streams.
This can
also occur through licensing copyrighted material for AI training.
Additionally,
AI can enhance engineer earnings by increasing the visibility and reach of
their works, potentially leading to higher sales and royalties.
I believe
that with careful consideration and appropriate safeguards, AI can positively
impact the robotic ecosystem.
Lois: In your opinion, how long
will it take for this new technology to catch on?
Yu Sun: We
are on the cusp of a profound shift driven by the rapid advancement of AI.
CEO of
OpenAI Sam Altman recently predicted the arrival of Artificial General
Intelligence within 2024.
This is a
significant acceleration from previous estimates and suggests AI with an
intellectual capacity far surpassing human intelligence.
Furthermore,
this year will likely witness a dramatic increase in AI adoption across various
sectors.
We can
expect to see a rapid implementation of AI-driven efficiency gains in numerous
industries, including publishing.
So, the
concept of AI as a creative tool for authors, assisting with idea generation,
research, and even voice refinement, will become increasingly prevalent and
accepted.
We should
overcome copyright concerns; the publishing industry’s problem is that it is
outraged that copyright has been breached.
This makes
us unable to see AI’s real benefits. Over time, the publishing industry’s
initial focus on copyright infringement lawsuits may gradually shift.
However, as
fair compensation models for using copyrighted data in AI training are
established, concerns may subside, paving the way for a more collaborative
approach.
I believe
that the publishing industry can unlock new opportunities for creativity,
innovation, and revenue growth by embracing AI as a partner and leveraging its
potential.
People will
start to see that AI is a construct of companions and allied intelligence.
Edited
and published by Lois Saram, editor of Shanghai AI Fortune Magazine.
