Towards Trusted AI Week 33 – Fool AI into ignoring stop signs

Secure AI Weekly admin todayAugust 23, 2021 59

share close

Machine learning and artificial intelligence can both be used for bad and for good


Fooled AI into ignoring stop signs using a cheap projector

The Next Web, August 16, 2021


A new adversarial attack can force the system to see or overlook what the malefactors need.

Purdue researchers presented their new study on adversarial attacks against сomputer vision systems. This new OPtical ADversarial attack (OPAD) can potentially affect many modern technologies that use computer vision, such as self-driving cars.

The new attack is a non-invasive one that uses structured illumination and relies on an inexpensive projector-camera system. During the attack, researchers project calculated templates to change the appearance of three-dimensional objects. In doing so, the main goal of the study becomes to identify potential dangers and then find ways to prevent them.

Deepfakes Are Now Making Business Pitches

Wired, August 16, 2021

Deepfake technologies make it possible to animate correspondence and presentations.

Several EY partners are currently testing a new technology based on artificial intelligence: it can animate customer presentations and even the simplest emails by voicing them in the style of a talking head. All this becomes possible with the help of these very deepfakes. All of this was done in order to simplify communication with customers during a pandemic, when live negotiations are difficult. The developers position their videos as fully synthetic and do not pursue the goal of deceiving their viewers.

“We’re using it as a differentiator and reinforcement of who the person is,” comments Jared Reeder, who works at EY. “As opposed to sending an email and saying ‘Hey we’re still on for Friday,’ you can see me and hear my voice”.

Now a consortium to prevent cyber attacks on Connected Cars

Cybersecurity Insiders, August 18, 2021

The goal of the new consortium will be to prevent hacking of autonomous vehicles and data theft.

Industry giants such as Toyota, Hyundai and Nissan have decided to form a consortium with the goal of combating attacks on unmanned vehicles. Microsoft Japan, Trend Micro, NTT Communication, Sompo Japan Insurance, Panasonic and Denso also said they are ready to support the initiative.

The problem of the safety of unmanned vehicles is becoming more and more acute in society due to the increasing demand for this type of transport. The popularity of such cars makes manufacturers increasingly think about their safety and security.

Written by: admin

Rate it
Previous post