Congress Tackles Deepfakes, Tensions Over Huawei Escalate, and a Major Data Breach at US CPB

Source: Deep Learning on Medium


This week in tech

The US Congress is (finally?) panicking about an AI technique colloquially called “deepfakes,” which is used to create fraudulent videos. The House of Representatives held its first hearing dedicated to the technology yesterday. During the hearing, Rep. Adam Schiff (D-CA) suggested repealing Section 230 of the Communications Decency Act, which gives social media platforms broad immunity from the law. In the meantime, Facebook published some improvements to the technology in question — specifically making advances in faking the speech of a target. Concerns about deepfakes are coupled with the fear that spam, trolling, and disinformation campaigns will increase in the run up to the 2020 presidential election.

Right on time, an example of deepfakes was published this week: a falsified video of Mark Zuckerberg was released on Instagram, where he says, “imagine this for a second: one man, in control of billions of people’s stolen data…” (video here). You could reasonably argue the video is satire, and you could reasonably argue the video is “fake news.” Both could be true depending on the context in which a viewer first discovers the video. For whatever it’s worth, Facebook has promised to leave the video up.

The concerns over disinformation is a major factor in the growing interest in regulating tech companies. On issues from privacy to antitrust to advertising discrimination to free speech and hate speech, politicians and government officials are looking to get a little more control over tech. Tech companies are responding by scaling up their lobbying efforts.

The heightening tensions between China and the US are also contributing to the tension between big tech and the U.S. government. Facebook banned Huawei from pre-installing Facebook on their new phones, and made the app harder to install in official ways on their devices. Last month Google made similar moves: in compliance with a government edict they shutdown access to the Play Store and other Android services for new Huawei phones. In response, Huawei has started building their own operating system. The Chinese Government has responded with threats of their own and Huawei, for it’s part, has invited security researchers to probe their systems for vulnerabilities.

Another major data breach was reported this week at U.S. Customs and Border Protection. Among the stolen data was biometric data such as images used for facial recognition. Security researchers are not at all surprised, and with the endless flood of breaches we shouldn’t be either. The latest attack is just one more reason we need better privacy legislation, and better data anonymizing tactics. One cool example of movement in the right direction is Apple’s new privacy preserving “Find My X” feature.

Other tidbits this week: The Chinese Government may have attacked messaging company Telegram to hamper the protestors in Hong Kong. More scientists are starting to say that drastic geoengineering might be necessary to stave off climate change, and Bitcoin is contributing more than you might think to that problem. The CCR5 knockout that Chinese geneticist He Jiankui made to two babies using CRISPR might significantly shorten the life expectancy of the children. And scientists were able to measure and reverse a “quantum leap,” which seems to be evidence that quantum leaps are actually a continuous process, rather than a discrete one.


This article is part of our weekly newsletter. Sign up for the Weekly Lab Report to have it delivered right to your inbox or visit our website to learn more.