Streaming Media
‘In the new era, thought itself will be transmitted by radio’
prophesied inventor Guglielmo Marconi and it seems telcos allied with Big
Tech are nearing that promise.
article here
It’s been three years since ChatGPT opened people’s eyes to
the potential of AI yet many telcos and their enterprise customers have
struggled with how to get a return on investment. The dial is now switching
from proof of concept to deployments at scale, according to executives at
Mobile World Congress in Barcelona.
“AI has passed the experimentation phase and now is the time
for industrialization,” declared Elena Gil Lizasoain, Director of Telefonica’s
AI & Data Business Unit. “That said, it is easier to do pilots than to
embed AI in real world operations. You need to combine LLMs with your own data
which is a process of change management, of training employees how to work with
new tech and of navigating ethical and privacy issues.”
KPMG’s Mark Gibson, a partner in its TMT division, shared
that contrary to popular belief, 76% of global CEOs believe AI will not reduce
headcount and that the overwhelming majority (92%) expect headcount to grow
with upskilling and redeployment.
“89% of the highest performing organisations say their
business is using AI to fill skill gaps among knowledge workers,” he said.
Results from the survey found 72% of business leaders
confident about the economic outlook that will come from digitization and
connectivity and that two thirds said they would continue to invest in AI
regardless of economic conditions.
The biggest impact
of AI at this stage will be in automating operations, both within customer
service including marketing and sales, and in the network.
“There are still big questions about how reliable the
technology is and how we deploy it in ways that drive actual business value,”
noted Lori Driscoll, PwC, Global TMT Consulting Leader.
Mitchell Gunnels, AVP Technology at AT&T explained that
AI is transforming his company to be digitally efficient. “We started with AI primitives - basic things
like copyright protections, a rock solid AI policy and abuse monitoring built
into the platform. Once we’d laid that foundation we could run applications on
the platform.”
The telco has trained its AI on 20 years of legacy code to
bring new software delivery requests forward from weeks to just 15-30 minutes.
“Beyond our internal IT we are now looking at external
revenue generation,” Gunnels said. “This area is exploding within AT&T.
“In order to get GenAI to do tasks for you, you need to
understand how to work with a co-intelligence,” he advised. “That’s something
we will all have to do. Don’t think
about a standard chatbot GUI. You need to embed AI in your workflows and your
networks.”
Enter the Agents
On the customer service side, a quick win for telcos is to
introduce AI-powered agents. So called
Agentic AI are persona-based agents that operate with a high degree of
autonomy.
“It is hard to understate the importance that this is the
year of agents,” said Calum Chace, CMO, Conscium.
In one sense agents have been around for a while. Alexa is
part of the furniture for many households, albeit one that is extremely simple,
Chace said. “It can’t reason a series of activities and carry them out without
human intervention. We can now create agents that do that. Everyone concerned
about AI safety is saying let’s not develop AI agents because once out in the
world we cannot control them. But it’s too late. The horse is half way out of
the stable.”
He said his company which he described as a “conscious AI
research organisation” can verify that agents “will do what want them to do and
will not go rogue.”
On the Agentic front, Chinese consumer electronics brand
Honor said it was investing U$D10 billion on transitioning into an “AI-first
device ecosystem company”, away from solely building smartphone hardware.
Orange, Telefonica and Vodafone have partnered with Honor in the shift. With
Google Cloud and Qualcomm Honor will launch “an intelligent smartphone” later
this year which will feature a personal AI agent capable of managing schedules,
making reservations, and optimising daily tasks using contextual awareness and
adapting to user behaviour.
Autonomous AI are “clearly the future” in the burgeoning
space industry, said Kimberly Washington, Co-Founder & CEO at Deep Space
Biology whose clients include NASA, but more was needed from a regulatory
perspective. “The space sector needs to establish a governance system but we
are not quite sure who owns the moon nor the planets beyond it,” she said.
Everything connected
In a conference session asking ‘Are Telcos Prepared for 35
Billion Devices by 2030?’ the consensus was that this figure was too low.
“There are already more people with mobile devices (8
billion) than with toothbrushes in the world today,” said Mikael Bäck of
Ericsson.
Noting that at two decades ago no-one in the telco industry
saw much business in everyone owning a smartphone, Ray Dolan, CEO of Cohere
Technologies said, “The number of business models that have collapsed because
of failing to bet on what we now take for granted means that we should be open
to disruption. But for innovation to scale there needs to be a common control
plane with a diversified user plane on top.”
One application riding a wave of mobile connectivity is
expected to be esports. According to Dario Betti, CEO of trade body Mobile
Ecosystem Forum, “Mobile gaming, already a dominant segment of the gaming
market, is increasingly becoming a cornerstone of the esports ecosystem.
“YouTube, Twitch, Huya and Trovo will continue to host
massive audiences for esports events, with mobile players and creators playing
a central role. This may drive demand for smartphones with superior cameras,
high-quality microphones, and robust editing software, as aspiring esports
players and streamers look to produce professional-grade content.”
Wearing a pair of Orion AR glasses Meta’s Rafael Camargo,
VP, Wearables Systems, Reality Lab, said that his company is on a mission to
merge the physical and digital world.
“Today our digital screens are isolated from the rest of
world. Yet our digital life is as important as our physical life. In Reality
Labs we work on how to merge them in ways that are familiar.”
An example: A realtime conversation between people who speak
and understand differently languages could be enabled by Meta’s AR headgear.
Another example: Conversations between two people in a crowded noisy ambient
environment could be enhanced to enable the two to chat without raising their
voices.
“The next question is how we develop social norms that
ensure you have agency and control. As we get more sensors in glasses it will
give you super powers close to your senses, your eyes and ears but society will
need standards around this. Should we, for example, enable facial recognition?
We could, but we don’t today. It will be a process of social norms and
etiquette about how to use the tech. That is a moving path and will take a
little while to establish.”
Going Quantum
Quantum computing will take AI to the next level but there
is some way to go yet. Nokia’s Mikael Rylander explained, “We are evolving a
new platform for digitisation in the world – which will be a mix of compute for
AI and traditional compute and communication. We need to be at the Edge and
also have central compute capability for rapid inferencing.
When we start to get the first Quantum compute online we
will plug that into this digital platform. we call Network Cloud Continuum.
Rylander caveated, “We must also realise that Quantum
compute right now is very error prone. A lot of work is going in to handle
error correction and get the qubits much more stable. Right now, Quantum can do
very advanced compute on very little data which is contrary to where we are
with AI.”
With 6G capabilities set to be released in 2028 we are
moving into a future where AI is intrinsic to network infrastructure. “AI will
be an integral component of network orchestration and management, signal
processing and network optimization, structurally changing how we think about and design our
communications systems,” said Andreas Roessler of Rohde & Schwarz.
In other words, there will be a shift from AI as a
performance enhancement to AI as a key technology component: from AI-assisted
to AI-native.
No comments:
Post a Comment