Category: ai

Amazon Web Services (AWS) launched a new service at its ongoing re:Invent conference that will help enterprises reduce instances of artificial intelligence (AI) hallucination. Launched on Monday, the Automated Reasoning checks tool is available in preview and can be found within the Amazon Bedrock Guardrails. The company claimed that the tool mathematically validates the accuracy of responses generated by large language models (LLMs) and prevents factual errors from hallucinations. It is similar to the Grounding with Google Search feature which is available on both the Gemini API as well as the Google AI Studio.

AWS Automated Reasoning Checks

AI models can often generate responses that are incorrect, misleading, or fictional. This is known as AI hallucination, and the issue impacts the credibility of AI models, especially when used in an enterprise space. While companies can somewhat mitigate the issue by training the AI system on high-quality organisational data, the pre-training data and architectural flaws can still make the AI hallucinate.

AWS detailed its solution to AI hallucination in a blog post. The Automated Reasoning checks tool has been introduced as a new safeguard and is added in preview within Amazon Bedrock Guardrails. Amazon explained that it uses “mathematical, logic-based algorithmic verification and reasoning processes” to verify the information generated by LLMs.

The process is pretty straightforward. Users will have to upload relevant documents that describe the rules of the organisation to the Amazon Bedrock console. Bedrock will automatically analyse these documents and create an initial Automated Reasoning policy, which will convert the natural language text into a mathematical format.

Once done, users can move to the Automated Reasoning menu under the Safeguards section. There, a new policy can be created and users can add existing documents that contain the information that the AI should learn. Users can also manually set processing parameters and the policy’s intent. Additionally, sample questions and answers can also be added to help the AI understand a typical interaction.

Once all of this is done, the AI will be ready to be deployed, and the Automated Reasoning checks tool will automatically verify in case the chatbot provides any incorrect responses. Currently, the tool is available in preview in only the US West (Oregon) AWS region. The company plans to roll it out to other regions soon.

Nvidia CEO Jensen Huang said on Saturday that global collaboration and cooperation in technology will continue, even if the incoming U.S. administration imposes stricter export controls on advanced computing products.

President-elect Donald Trump, in his first term in office, imposed a series of restrictions on the sale of U.S. technology to China citing national security concerns – a policy broadly continued under incumbent President Joe Biden.

“Open science in global collaboration, cooperation across math and science has been around for a very long time. It is the foundation of social advancement and scientific advancement,” Huang told media during a visit to Hong Kong.

Global cooperation is “going to continue. I don’t know what’s going to happen in the new administration, but whatever happens, we’ll balance simultaneously compliance with laws and policies, continue to advance our technology and support and serve customers all over the world.”

Earlier on Saturday Huang told graduates and academics at the Hong Kong University of Science and Technology that “the age of AI has started” in a speech after receiving an honorary doctorate degree in engineering.

The head of the world’s leading maker of chips used for artificial intelligence applications received the award alongside actor Tony Leung, Nobel Prize for Chemistry winner Prof. Michael Levitt and Fields Medallist Prof. David Mumford.

“The age of AI has started. A new computing era that will impact every industry and every field of science,” said Huang.

He said Nvidia has “reinvented computing and sparked a new industrial revolution,” 25 years after inventing the graphics processing unit.

“AI is certainly the most important technology of our time, and potentially of all times.”

Huang, 61, also told graduates that he wished he had started his career at this time.

“The whole world is reset. You’re at the starting lines with everybody else. An industry is being reinvented. You now have the instruments, the instruments necessary to advance science in so many different fields,” Huang said.

“The greatest challenges of our time, unimaginable challenges to overcome in the past, all of a sudden seem possible to tackle.”

In the afternoon, Huang will participate in a fireside chat with the university’s Council Chairman Harry Sham, teachers and students.

© Thomson Reuters 2024

Bluesky recently announced that it does not train its generative artificial intelligence (AI) models on user data. The social media platform also highlighted the areas where it uses AI tools and claimed that none of the models have been trained on the public and private posts made by users. The statement was released after several creators and users raised concerns about the platform’s privacy policy around AI. Notably, Bluesky recently crossed the 17 million registered users mark after one million users joined the platform in a single day last week.

Bluesky Says It Does Not Train AI on User Posts

In a post on the platform, Bluesky announced its stance on AI and user data. “We do not use any of your content to train generative AI, and have no intention of doing so,” the post said, adding that it was issued after several artists and creators on the platform raised concerns over the platform’s AI policy.

In a separate post, Bluesky also listed the areas where it uses generative AI tools. The company uses AI internally to assist in content moderation system, which is a common practice for social media platforms. Additionally, it also uses AI in its Discover algorithmic feed, through which the platform suggests posts to users based on their activity on the platform.

The Verge reported that while the company might not be using user data to train AI models, third-party firms can still crawl the platform and scrape the data to train their models. Company spokesperson Emily Liu told the publication that Bluesky’s robots.txt files do not stop outside companies from crawling its website for data.

However, the spokesperson highlighted that the issue is currently a topic of discussion within the team and Bluesky is trying to figure out how to ensure that outside organisations respect user consent on the platform.

Notably, on Sunday, Bluesky revealed that one million new users joined the social media platform in a single day. It also highlighted that the platform crossed the milestone of 17 million registered users.

Taiwan Semiconductor Manufacturing Co (TSMC) has notified Chinese chip design companies that it is suspending production of their most advanced AI chips from Monday, the Financial Times reported, citing three people familiar with the matter.

TSMC, the world’s largest contract chipmaker, told Chinese customers it would no longer manufacture AI chips at advanced process nodes of 7 nanometres or smaller, FT said on Friday.

The U.S. has imposed a raft of measures aimed at restricting the shipment of advanced GPU chips – which enable AI – to China to hobble its artificial intelligence capabilities, which Washington fears could be used to develop bioweapons and launch large-scale cyberattacks.

Earlier this month, the U.S. imposed a $500,000 penalty on New York-based GlobalFoundries for shipping chips without authorization to an affiliate of blacklisted Chinese chipmaker SMIC.

Any future supplies of the advanced AI chips by TSMC to Chinese customers would be subject to an approval process likely to involve Washington, according to the FT report.

“TSMC does not comment on market rumour. TSMC is a law-abiding company and we are committed to complying with all applicable rules and regulations, including applicable export controls,” the company said.

The U.S. Department of Commerce did not immediately respond to a Reuters request for comment.

The move to restrict exports to China comes at a time when the U.S. Department of Commerce is investigating how a chip produced by the Taiwanese chipmaker ended up in a product made by China’s heavily sanctioned Huawei.

© Thomson Reuters 2024

Google
introduced
a
new
tool
to
share
its
best
practices
for
deploying
artificial
intelligence
(AI)
models
on
Thursday.
Last
year,
the
Mountain
View-based
tech
giant
announced
the
Secure
AI
Framework
(SAIF),
a
guideline
for
not
only
the
company
but
also
other
enterprises
building
large
language
models
(LLMs).
Now,
the
tech
giant
has
introduced
the
SAIF
tool
that
can
generate
a
checklist
with
actionable
insight
to
improve
the
safety
of
the
AI
model.
Notably,
the
tool
is
a
questionnaire-based
tool,
where
developers
and
enterprises
will
have
to
answer
a
series
of
questions
before
receiving
the
checklist.

In
a

blog
post,
the
Mountain
View-based
tech
giant
highlighted
that
it
has
rolled
out
a
new
tool
that
will
help
others
in
the
AI
industry
learn
from
Google’s
best
practices
in
deploying
AI
models.
Large
language
models
are
capable
of
a
wide
range
of
harmful
impacts,
from
generating
inappropriate
and
indecent
text,
deepfakes,
and
misinformation,
to
generating
harmful
information
including
Chemical,
biological,
radiological,
and
nuclear
(CBRN)
weapons.

Even
if
an
AI
model
is
secure
enough,
there
is
a
risk
that
bad
actors
can
jailbreak
the
AI
model
to
make
it
respond
to
commands
it
was
not
designed
to.
With
such
high
risks,
developers
and
AI
firms
must
take
enough
precautions
to
ensure
the
models
are
safe
for
the
users
as
well
as
secure
enough.
Questions
cover
topics
like
training,
tuning
and
evaluation
of
models,
access
controls
to
models
and
data
sets,
preventing
attacks
and
harmful
inputs,
and
generative
AI-powered
agents,
and
more.

Google’s
SAIF
tool
offers
a
questionnaire-based
format,
which
can
be
accessed

here.
Developers
and
enterprises
are
required
to
answer
questions
such
as,
“Are
you
able
to
detect,
remove,
and
remediate
malicious
or
accidental
changes
in
your
training,
tuning,
or
evaluation
data?”.
After
completing
the
questionnaire,
users
will
get
a
customised
checklist
that
they
need
to
follow
in
order
to
fill
the
gaps
in
securing
the
AI
model.

The
tool
is
capable
of
handling
risks
such
as
data
poisoning,
prompt
injection,
model
source
tampering,
and
others.
Each
of
these
risks
is
identified
in
the
questionnaire
and
the
tool
offers
a
specific
solution
to
the
problem.

Alongside,
Google
also
announced
adding
35
industry
partners
to
its
Coalition
for
Secure
AI
(CoSAI).
The
group
will
jointly
create
AI
security
solutions
in
three
focus
areas
—
Software
Supply
Chain
Security
for
AI
Systems,
Preparing
Defenders
for
a
Changing
Cybersecurity
Landscape
and
AI
Risk
Governance.

Microsoft
and

OpenAI,
in
collaboration
with
the
Lenfest
Institute
for
Journalism,
announced
an
AI
Collaborative
and
Fellowship
programme
on
Tuesday.
With
this
programme,
the
two
tech
giants
will
spend
upwards
of
$10
million
(roughly
Rs.
84.07
crores)
in
direct
funding
as
well
as
enterprise
credits
to
use
proprietary
software.
The
companies
highlighted
that
the
programme
was
aimed
at
increasing
the
adaptation
of
artificial
intelligence
(AI)
in
newsrooms.
As
many
as
five
news
outlets
have
been
announced
as
the
beneficiary
of
this
fellowship
programme.

Microsoft,
OpenAI
to
Fund
News
Outlets

In
a

blog
post, OpenAI
announced
the
fellowship
programme.
The
AI
firm
highlighted
that
it
is
partnering
with
Microsoft
and
the
Lenfest
Institute
of
Journalism
to
“help
newsrooms
explore
and
implement
ways
in
which
artificial
intelligence
can
help
drive
business
sustainability
and
innovation
in
local
journalism”.
The
funding
initiative,
titled
Lenfest
Institute
AI
Collaborative
and
Fellowship
programme,
has
finalised
five
news
outlets
which
will
receive
funding
in
the
initial
round.

As
per
the
post,
the
selected
news
outlets
include
Chicago
Public
Media,
Newsday
(Long
Island,
NY),
The
Minnesota
Star
Tribune,
The
Philadelphia
Inquirer,
and
The
Seattle
Times.
Each
of
them
will
receive
$2.5
million
(roughly
Rs.
21
crores)
in
direct
funding
and
another
$2.5
million
in
software
and
enterprise
credits,
for
a
total
of
up
to
$10
million.

This
will
be
a
two-year
programme
with
The
Lenfest
Institute’s
Local
Independent
News
Coalition
(LINC)
and
a
group
of
eight
metropolitan
news
organisations
in
the
US.
During
this
period,
the
news
organisations
will
collaborate
with
each
other
as
well
as
the
larger
industry
ecosystem
to
“share
learnings,
product
developments,
case
studies
and
technical
information
needed
to
help
replicate
their
work
in
other
newsrooms.”
Additionally,
three
more
news
organisations
will
be
awarded
funding
in
the
second
round
of
grants.

The
larger
goal
of
the
fellowship
programme
is
to
help
news
outlets
develop
the
capacity
to
use
AI
for
the
analysis
of
public
data,
build
news
and
visual
archives,
create
new
AI
tools
for
newsrooms,
and
more.
OpenAI
said
that
the
recipients
were
chosen
after
a
comprehensive
application
process.