Digimasters Shorts - AI Fails Journalism, Elon Musk’s Grok Leaks User Chats, Microsoft Debuts Copilot 3D, Battery AI Revolution, Apple Eyes Google Gemini for Siri

Send us a text

Welcome to Digimasters Shorts, your quick dive into the latest developments in the digital world. Hosted by Adam Nagus and Carly Wilson, each episode brings you sharp insights on AI's role and limitations in journalism, privacy concerns surrounding generative AI tools like Grok, Microsoft's latest innovations in 3D modeling with Copilot, breakthroughs in battery technology driven by AI, and the evolving AI strategies of tech giants like Apple and Google. Stay ahead of the curve with concise updates and expert analysis on how these tech trends are shaping our future.

Support the show

Don't forget to checkout our larger sister podcast - The Digimasters Podcast here. Which has many expert guests discussing AI, Career Mentoring, Fractional Careers, Digital and much much more.

Episode Transcript

Adam N2 (00:05):
Welcome to Digimasters Shorts, we are your hosts Adam
Nagus

Carly W (00:09):
and Carly Wilson delivering the latest scoop from
the digital realm.
A recent Columbia JournalismReview study challenges the
notion that AI is an effectiveresearch assistant for
journalists.
Tested tools, including ChatG.P.T and Claude, struggled with
specific tasks like summarizinglocal government meeting
transcripts and conductingscientific literature reviews.

(00:29):
While short summaries performedreasonably well, longer
summaries often containedinaccuracies and missed facts.
Research tools fared worse,frequently citing irrelevant or
inconsistent papers.
The study highlights the needfor careful prompt crafting and
iterative evaluation to improveAI output.
It also reveals that AIapplications oversimplify
controls, limiting users’ability to tailor results

(00:52):
effectively.
Contrary to some beliefs, AIliteracy remains crucial for
journalists to maximize thesetools’ benefits.
Newsrooms are encouraged tosystematically test AI
internally with dedicated teams,fostering an iterative learning
process.
Ultimately, the choice for manyjournalists is between imperfect
AI assistance and no assistanceat all.

(01:14):
Incremental improvements throughtraining and experimentation may
gradually raise the quality andvalue of AI support in
journalism.

Adam N2 (01:21):
Elon Musk's AI company xAI has launched Grok, a
generative AI similar to ChatG.P.T.
However, Grok has raised privacyconcerns after over 370,000 user
chats appeared on Google’ssearch results.
When users press the“share”button, their conversations are
published on Grok’s websitewithout clear warnings.
These shared chats generateunique URLs that are indexed by

(01:43):
search engines like Google,Bing, and DuckDuckGo.
Though the chats aren’t directlytied to usernames, sensitive
information in transcripts couldidentify users indirectly.
Conversations revealed rangedfrom mundane business tasks to
disturbing topics like hackingattempts and fictional terrorist
attacks.
Some users even disclosedintimate details, including

(02:03):
personal names and passwords.
This exposure has sparkedconcern about users’ privacy and
data security.
Grok is not alone; Open A.I’sChat G.P.T faced similar issues
earlier this month and had todisable its sharing feature amid
public backlash.
The incident highlights ongoingchallenges in balancing AI
innovation with user privacysafeguards.

(02:24):
Microsoft has expanded its AIassistant capabilities with the
latest launch of Copilot 3D, atool that turns 2D images into
3D models.
Integrated into its extensiveuser base of over 1.4 billion
Windows devices, Microsoft aimsto strengthen its position
against competitors like ChatG.P.T.
Copilot already supports varioustasks, from drafting emails to

(02:46):
summarizing documents and evenassisting in video games.
The service has introducedfeatures such as AI-powered
podcast creation and a memorycomponent that personalizes user
interaction.
To create a 3D model, usersupload JPG or PNG images up to
10MB on the Copilot 3D site, andthe conversion takes under a
minute.

(03:07):
The generated models are savedfor 28 days under users'
accounts and can be downloadedin GLB format for further use.
Microsoft recommends images witha single subject, even lighting,
and plain backgrounds foroptimal results.
This addition follows recentpricing changes linked to
integrating Copilot intoMicrosoft 365.

(03:27):
Tens of millions of users haveadopted Microsoft Copilot as of
May this year.
With these advancements,Microsoft seeks to enhance daily
productivity and appeal tocreators across sectors.

Carly W (03:38):
The global race for better batteries is more intense
than ever, driven by demand fromelectric vehicles, drones, and
next-generation aircraft.
Traditional battery R&Dstruggles with lengthy testing
cycles, often taking up to eightmonths to prove 500
charge-discharge cycles.
However, physics-informed AI isrevolutionizing this process by

(03:58):
embedding physical laws intomodels, enabling predictions of
battery life up to 1,000 timesfaster than conventional
methods.
National labs and companies likeFactorial have demonstrated that
early testing of just one to twoweeks can now forecast long-term
battery performance.
Factorial’s Gammatron platformrecently doubled battery cycle
life through software-drivenoptimization without hardware

(04:20):
changes.
Similar advances at companieslike Monolith are reducing
testing times by up to 70%,accelerating product launch
timelines by months.
This marks a shift fromhardware-first to data-first
innovation, combiningsimulations, lab data, and AI to
redefine battery development.
The adoption of these advancedAI tools will separate industry

(04:41):
leaders from followers byenabling faster charging, longer
range, and increased resilience.
The battery industry stands onthe brink of a digital design
revolution, where breakthroughsstart in virtual labs before
physical prototypes.
The speed of adoption willdetermine who leads in the next
generation of energy storagetechnology.
Apple is reportedly in talks tointegrate Google's Gemini

(05:03):
chatbot into its iOS assistant,Siri, aiming for a potential
2026 release.
According to sources, Apple hasapproached Alphabet about a
custom model, with Googletesting versions that can
operate on Apple’s servers.
This move follows months ofApple exploring external
partnerships to enhance its AIcapabilities.
Earlier rumors suggested Applemight abandon its in-house Siri

(05:26):
AI project, known internally as"Linwood," in favor of external
models named"Glenwood." Thecompany considered
collaborations with Open A.I’sChat G.P.T and Anthropic’s
Claude before focusing onGoogle’s Gemini.
Industry insiders see this shiftas part of Apple's strategy amid
AI talent losses and changingmarket dynamics.
Other tech giants have recentlyscaled back AI investments,

(05:49):
indicating a possible coolingoff in the AI boom.
While neither Apple nor Googlehave confirmed the partnership,
both have hinted that Geminiintegration could complement
their existing collaboration onSearch.
This potential deal mirrorsprevious arrangements Apple made
to incorporate Chat G.P.Tfeatures.
If realized, the update couldsignificantly enhance Siri's AI

(06:09):
functionality and userexperience.

Don (06:12):
Thank you for listening to today's AI and Tech News podcast
summary...
Please do leave us a comment andfor additional feedback, please
email us [e-mail] You
can now follow us on Instagramand Threads by searching for
@DigimastersShorts or Search forDigimasters on Linkedin.
Be sure to tune in tomorrow anddon't forget to follow or
subscribe!