Category: Trust
Top AI researchers race to detect ‘deepfake’ videos: ‘We are outgunned’ | THE WASHINGTON POST
The ability to create synthetic videos outweighs detection methods by ‘100 to 1’, according to a digital forensics professor at UC Berkeley quoted in THE WASHINGTON POST. The US public is not ready for the possible consequences of deepfakes in…
Leading the way into the age of artificial intelligence | FINLAND 2019
THE GOVERNMENT OF FINLAND wants AI to be ‘human-centric’ and ‘trust-generating.’ The aspiration comes in a sweeping government paper on national readiness, saying AI is more than a technology. AI understanding should reach all aspects of Finnish society. It should…
You can train an AI to fake UN speeches in just 13 hours | MIT TECHNOLOGY REVIEW
‘The researchers conclude that a greater global effort is needed to work on ways of detecting and responding to AI-generated content.’ With $7.80 in computer time and 13 hours of human effort, two UN researcher coaxed an AI system to…
If we are using AI in journalism we need better guidelines on reporting uncertainty | ONLINE JOURNALISM BLOG
‘85% is quite good in machine learning terms — but journalistically, it’s problematic. We have moved from dealing with facts, to dealing with estimates.’ The ONLINE JOURNALISM BLOG points out differences between reporting data as facts and data as probabilities…
Five questions you can use to cut through AI hype | MIT TECHNOLOGY REVIEW
Oftentimes, companies pass the first four tests but not the last. For me, that is a major red flag. MIT TECHNOLOGY REVIEW offers a checklist by Karen Ho to help assess hype-heavy claims about products ‘powered’ by artificial intelligence. Her…
Ethics guidelines for trustworthy AI | EUROPEAN COMMISSION 2018
The EUROPEAN COMMISSION presents seven guidelines for trustworthy artificial intelligence. The Commission says they want to create ‘an environment of trust’ when developing AI as well as when deploying and using AI. Each of the seven headings links to the…
Mirroring to Build Trust in Digital Assistants | Metcalf et al
The chattiness of an AI-generated voice impacts how much it’s trusted and varies by user, according to researchers at Apple. They found useability increased when the conversational style of a voice assistant more closely matched individual preferences. They suggest conversational…
How Reuters is training reporters to spot ‘deepfakes’ | DIGIDAY
DIGIDAY UK reports how Reuters created its own fake video to test its in-house verification team and step-up its ability to detect deepfakes. ‘In the last two years, Reuters has doubled the number of people who work on verifying video…
Trust in Public Policy Algorithms | Kennedy et al
‘We find a stunning degree of trust in algorithms relative to other sources across all of the experiments.’ There is an ‘urgent need’ for more public awareness of algorithms reports this research team after conducting six studies of public attitudes….
New AI fake text generator may be too dangerous to release, say creators | THE GUARDIAN
OpenAI says it’s withholding some of the science behind its text-generating AI model GPT-2. THE GUARDIAN reports the company is concerned that the believability of GPT-2’s text could be used for public deception. Possibilities include fake reviews, fake news, and…
This is how AI bias really happens—and why it’s so hard to fix | MIT TECHNOLOGY REVIEW
‘Bias can creep in at many stages of the deep-learning process, and the standard practices in computer science aren’t designed to detect it.’ MIT TECHNOLOGY REVIEW provides a helpful primer on the ways bias enters AI systems. POSSIBLE SOURCES OF…
What kind of news gatekeepers do we want machines to be? Filter bubbles, fragmentation, and the normative dimensions of algorithmic recommendations | Nechushtai & Lewis
Efrat Nechushtai and Seth Lewis burst filter bubbles in this paper about choices made by news algorithms. They researched automated story selection by Google News in the 2016 US Presidential election. They found liberals and conservatives were shown the same…
WaPo Editorial: A reason to despair about the digital future – Deepfakes | THE WASHINGTON POST
‘The ability to use machine learning to simulate an individual saying or doing almost anything poses personal and political risks that societies around the world are ill-equipped to guard against.’ IN THE WASHINGTON POST the editorial board warns that algorithm-generated…
How to recognize fake AI-generated images | MEDIUM
‘Here are some things you can look for when trying to recognize an image produced by a GAN.’ MEDIUM provides several examples of ‘tells’ for fake imagery. They include incongruous backgrounds, asymmetrical body parts, peculiar hair treatments, and others. SEE…
Trust highest in Sweden and Germany, lowest in Hungary and U.K. | 9 country study
Researchers at Cambridge University teamed with YouGov to understand attitudes about trust in media. The study was conducted in summer 2018 and included favoured media sources and attitudes to government. Journalistic trust levels in Sweden and Germany were each at…
How The Wall Street Journal is preparing its journalists to detect deepfakes | NEIMANLABS
“Raising awareness in the newsroom about the latest technology is critical,” said Christine Glancey, a deputy editor on the Ethics & Standards team who spearheaded the forensics committee. “We don’t know where future deepfakes might surface so we want all eyes…
Do People Trust Algorithms More Than Companies Realize? | HBR
‘…it may be not be necessary to invest in emphasizing the human element of a process wholly or partially driven by algorithms.’ Research published in THE HARVARD BUSINESS REVIEW says people favour advice from algorithms over people giving the same…
Marcus & Ernest: No, artificial intelligence won’t solve fake news | NEW YORK TIMES
‘Existing A.I. systems that have been built to comprehend news accounts are extremely limited… lacking a robust mechanism for drawing inferences or a way of connecting to a body of broader knowledge.’ IN THE NEW YORK TIMES Gary Marcus and…
It’s Getting Harder to Spot a Deep Fake Video | BLOOMBERG
Runs 2:59 Report by Bloomberg News.
Joseph Stiglitz: ‘going towards a more divided society’ | THE GUARDIAN
‘Which is the easier way to make a buck: figuring out a better way to exploit somebody, or making a better product? With the new AI, it looks like the answer is finding a better way to exploit somebody.’ IN…
Tom Kent: Fake news is about to get so much more dangerous | THE WASHINGTON POST
‘Society now has to learn that video no longer guarantees reliability. Instead, it could be the biggest lie of all.’ IN THE WASHINGTON POST Tom Kent sounds an alarm for media companies to increase care with 3rd-party video. His warning…
Oscar Schwartz: ‘The discourse is unhinged: how the media gets AI alarmingly wrong’ | THE GUARDIAN
‘Social media has allowed self-proclaimed ‘AI influencers’ who do nothing more than paraphrase Elon Musk to cash in on this hype with low-quality pieces. The result is dangerous.’ IN THE GUARDIAN Oscar Schwartz says trivializations of AI in media coverage…
What to Watch for in the Coming Wave of “Deep Fake” Videos | GIJN
…“deep fake” videos can be surprisingly convincing, and the software to make them has emerged rapidly’ As algorithms for image and audio processing become more powerful, so does the ability to use them for deceit. Manipulation techniques applied to video…
Public Trust in Journalism: An Annotated Bibliography | CMT 2018
THE CENTRE FOR MEDIA TRANSITION offers this annotated bibliography about public trust in journalism. More than 70 pages present a compendium of annotated sources from around the world, including: Studies of public attitudes Peer-reviewed academic papers Journalistic accounts Active links…
Giving Computers a Nose for News: Exploring the limits of story detection and verification | Thurman et al
Using algorithms for story discovery heralds a new phase for social media says Neil Thurman and his co-authors. They look at the implications when algorithms assess social posts and take a detailed look at one in particular, SocialSensor. Their study…
Algorithmic Accountability | Diakopoulos
Nick Diakopoulos examines the power of algorithms. He considers how they function, their influence in society, and suggests approaches to increasing accountability. His paper analyzes five cases in journalism. He is Director of the Computational Journalism Lab at Northwestern University,…