/
Navigation
C
Chronicles
Browse all articles
C
E
Explore
Semantic exploration
E
R
Research
Entity momentum
R
N
Nexus
Correlations & relationships
N
~
Story Arc
Topic evolution
S
Drift Map
Semantic trajectory animation
D
P
Posts
Analysis & commentary
P
Browse
@
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
?
Concept Search
Semantic similarity search
!
High Impact Stories
Top coverage by position
+
Sentiment Analysis
Positive/negative coverage
*
Anomaly Detection
Unusual coverage patterns
Analysis
vs
Rivalry Report
Compare two entities head-to-head
/\
Semantic Pivots
Narrative discontinuities
!!
Crisis Response
Event recovery patterns
Connected
Nav: C E R N
Search: /
Command: ⌘K
Embeddings: large
VOICE ARCHIVE

Sam Gregory

@samgregory
15 posts
2024-05-21
“Delightfully deceiving voters doesn't excuse the deception” I talked to @NilChristopher for this great deep-dive into the realities of and the people behind... 50 million + AI-clone voice calls and more... in Indian politics. https://www.wired.com/...
2024-05-21 View on X
Wired

How Indian politicians are using deepfakes for voter outreach; an estimated 50M+ AI voice calls were made in the two months before the elections in April 2024

India's elections are a glimpse of the AI-driven future of democracy.  Politicians are using audio and video deepfakes of themselves …

2024-01-27
Easier production of fake audio, no shared classifiers, no watermarking standards, loose and inadequate company guardrails and... as I noted to @Knibbs critical gap in access to detection skills/tools for frontline journalists/election officials (in US, even + globally)
2024-01-27 View on X
Bloomberg

Source: ElevenLabs banned a creator's account after researchers said the company's tech was used to make an audio deepfake used in a fake Biden robocall

Company's technology used to create fake Biden recording, researchers say  —  The creator of an audio deepfake of US President Joe Biden urging people …

2023-09-30
AI-generated audio playing a role in both disinformation and political ads in Slovakia elections. Not a surprise, audio deepfakes - PLUS claims that real compromising audio was faked as we see in @witnessorg cases- likely to reoccur in upcoming elections. https://www.bloomberg.com/...
2023-09-30 View on X
Bloomberg

Research: ahead of Slovakia's parliamentary elections, AI-generated deepfake videos of politicians are spreading on Facebook, Telegram, and other platforms

Videos featuring AI-generated deepfake voices of politicians are spreading on social media ahead of the Slovak parliamentary elections this weekend …

2023-07-31
With provenance in headlines, good look @C2PA_org from @TateRyMo @techreview. Another dimension: how these types of provenance tools must be resilient in global/diverse settings, raising privacy risks, access questions, implied truth effects and threats of abusive legislation
2023-07-31 View on X
MIT Technology Review

A look at C2PA, a standard that relies on cryptography to encode provenance information of media content, started by Adobe, Arm, Intel, Microsoft, and Truepic

An internet protocol called C2PA adds a “nutrition label” to images, video, and audio.  —  The White House wants big AI companies …

2023-07-24
Watermarking and provenance to show when something is AI-made audiovisual is one commitment made by tech companies to White House. Let's take close read of that para of White House-AI companies commitments on trust, watermarking and provenance ( https://www.whitehouse.gov/... [image]
2023-07-24 View on X
New York Times

An analysis of the potential impact of the eight AI commitments to the White House made by OpenAI, Microsoft, Meta, Google, Amazon, Anthropic, and Inflection

2023-07-23
Watermarking and provenance to show when something is AI-made audiovisual is one commitment made by tech companies to White House. Let's take close read of that para of White House-AI companies commitments on trust, watermarking and provenance ( https://www.whitehouse.gov/... [image]
2023-07-23 View on X
New York Times

An analysis of eight AI commitments made by OpenAI, Microsoft, Meta, Google, Amazon, Anthropic, and Inflection to the White House, and their potential impact

Seven leading A.I. companies made eight promises about what they'll do with their technology.  Our columnist sizes up their potential impact.

2023-05-01
A foundational figure in deep learning @geoffreyhinton critiques the current headlong #generativeai rush as resigns from Google, starting with the flood of false photos, videos and text and inability to trust... https://www.nytimes.com/... https://twitter.com/...
2023-05-01 View on X
New York Times

An interview with AI pioneer Geoffrey Hinton, who recently left Google after over a decade to speak about AI's risks, on regrets, quitting the company, and more

For half a century, Geoffrey Hinton nurtured the technology at the heart of chatbots like ChatGPT.  Now he worries it will cause serious harm.

2023-04-30
Couldn't agree more with quote in great @joannastern explainer: “you can't make everyone an AI detective” @witnessorg consultations say don't devolve responsibility to every individual, and that its getting easier to “carpet-bomb the internet with it” https://www.wsj.com/...
2023-04-30 View on X
Wall Street Journal

A WSJ columnist used Synthesia, ElevenLabs, and ChatGPT to create an AI video clone that could trick her bank's voice biometric system, make a TikTok, and more

This is a perfect example in the difference between proper #ai technology … Tweets: @yogapantschik : Well this is just fabulous! India, Philippines & Nigeria may actually convince ...

2023-04-29
Couldn't agree more with quote in great @joannastern explainer: “you can't make everyone an AI detective” @witnessorg consultations say don't devolve responsibility to every individual, and that its getting easier to “carpet-bomb the internet with it” https://www.wsj.com/...
2023-04-29 View on X
Wall Street Journal

A WSJ columnist used Synthesia, ElevenLabs, and ChatGPT to create an AI clone that could trick her bank's voice biometric system, make a TikTok, and more

This is a perfect example in the difference between proper #ai technology … Jay Meier : Voice biometrics has never really been shown to be terribly effective.  I know banks have us...

2023-04-08
And irrespective of whether Meta sends AI-generated images to fact-checkers (challenge: it's slower and they lack skills/tech on AI-images), or uses automation in-house (challenge: detection is faulty)... Houston we have a problem either way if volume of AI-generated images ↗️↗️
2023-04-08 View on X
Washington Post

Sources: Meta told political campaigns in March that the company reviews AI-generated images via its fact-checking program, which exempts candidates and parties

Isaac Stanley-Becker / Washington Post :

2023-03-27
The six-fingered hand clue won't last (like the ‘no blinking’ tip didn't). Lesson #1 of #deepfake detection - the visual Achilles heel to spot them goes away quickly... Better to build good/available authenticity/detection tools, and media literacy https://www.wired.com/... https://twitter.com/...
2023-03-27 View on X
Washington Post

A mid-March Midjourney update seemingly fixed one of the AI image generator's major failings: its inability to depict lifelike human hands with five fingers

Pranshu Verma / Washington Post :

2021-07-19
Glad @hels asked for my responses to backlash around AI-generated lines in new Bourdain documentary: we've been focused on these issues (and work-shopped draft norms) as part of our @witnessorg 'Prepare, Don't Panic' initiative on #deepfakes (https://wit.to/... to learn +) https://twitter.com/...
2021-07-19 View on X
New Yorker

Experts say consent and disclosure are key to the ethical use of synthetic media, amid controversy surrounding Bourdain's deepfake voice in a new documentary

2021-07-16
Faking Bourdain's voice? Is it appropriate to deepfake a dead person's voice? Under what circumstances and with what disclosure? We've been working on ethics principles around deepfakes @witnessorg. Consent, deceptiveness, disclosure, audience expectations in genre all key.. https://twitter.com/...
2021-07-16 View on X
The Verge

The maker of a new documentary about Anthony Bourdain says he used AI to re-create Bourdain's voice and synthesized the audio of three quotes from the TV host

‘We can have a documentary-ethics panel about it later’  —  In a new documentary, Roadrunner, about the life and tragic death …

2021-03-06
For everyone who asked me if should be worried uploading great-grandma's photo to the ‘resurrect them’ app Deep Nostalgia because privacy-compromising (of many reasons to potentially be creeeped out)... You're looking in wrong direction on scale of images being used to train AI. https://twitter.com/...
2021-03-06 View on X
OneZero

Facebook's AI model SEER was designed to exclude Instagram images from the EU in its dataset, likely to avoid GDPR violations

Dave Gershgorn / OneZero : Tweets: @histoftech , @samgregory , @kimcrayton1 , and @profcarroll Tweets: Mar Hicks / @histoftech : Facebook's earliest iteration, Facemash, also scra...

2019-09-25
+1 #deepfakes dataset to enhance detection - this time via @GoogleAI. We need coordination + sharing of datasets (esp once #deepfakes are + widespread) and real attention on how ensure detection methods= accessible, explainable, relevant to real-world + who is harmed by deepfakes https://twitter.com/...
2019-09-25 View on X
Google AI Blog

Google has released a dataset of ~3,000 deepfake videos to help researchers develop deepfake detection techniques

Deep learning has given rise to technologies that would have been thought impossible only a handful of years ago.  Modern generative models are one example of these …