OpenAI withholds full GPT-2 model over misuse concerns
14 Feb 2019
OpenAI unveils GPT-2, a text generator it says is too open to misuse for a full release, sparking debate about responsible AI publication.
Headline-making news in information security, cyber security, privacy, and AI, from 2010 to today.
OpenAI unveils GPT-2, a text generator it says is too open to misuse for a full release, sparking debate about responsible AI publication.
Singapore’s Ministry of Health says confidential details of 14,200 people diagnosed with HIV were leaked online by a foreign national.
France’s CNIL fines Google €50 million, then the largest GDPR penalty, for unclear information and invalid consent for ad personalisation.
Troy Hunt reveals Collection #1, a huge set of breached credentials holding around 773 million unique email addresses and 21 million unique passwords.
Marriott disclosed that attackers had access to its Starwood guest reservation database since 2014, exposing data on up to 500 million guests.
The Netherlands revealed it had caught four Russian military intelligence officers trying to hack the OPCW in The Hague.
Facebook said attackers exploited its ‘View As’ feature to steal access tokens, later putting the number affected at 29 million.
BA said attackers stole payment card details from its website and app, leading to a £20m ICO fine in 2020.
Attackers stole personal data of 1.5 million SingHealth patients, including Prime Minister Lee Hsien Loong.
A ransomware attack disrupted Atlanta’s city services for weeks, with recovery costs reported in the millions of dollars.
An Uber self-driving test car struck and killed a woman in Tempe, Arizona, the first pedestrian death involving an autonomous car.
Reports revealed that data on tens of millions of Facebook users had been collected through a quiz app and used for political targeting.