end-to-end ai smart driving is emerging, and safety issues are attracting much attention
2024-09-07
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
these days, the 27th chengdu international auto show is in full swing. many car companies have exhibited their new intelligent models equipped with ai big models. among them, the end-to-end ai big model has become the new technological highlight of this auto show, demonstrating a new height of automotive intelligence development. it also means that intelligent driving technology is being rapidly iterated and optimized at an unprecedented speed.
with the development of science and technology, the vulnerability and potential security risks of intelligent systems have also come along. in fact, there are two main common types of cyber attacks on smart connected cars: in-vehicle network attacks and out-vehicle network attacks. relevant data show that with the rapid development of large model technology, related cyber attacks are also increasing. the "2024 semi-annual automotive vulnerabilities and threat intelligence" released by a smart car security company shows that 70% of the security incidents that occurred worldwide in the first half of the year were remote attacks, especially smart cockpits, charging services, smart car control, autonomous driving, etc., which are all scenarios with high security threats. user privacy security faces severe challenges.
common methods of cyber attacks on smart connected cars
in this context, china automotive evaluation, the "national team of automobile evaluation", has actively responded to consumer concerns and industry calls, launched the first privacy protection evaluation project in the newly upgraded china intelligent connected vehicle technical specification (c-icap), and introduced attack defense guards to help automakers effectively improve the network attack defense capabilities of intelligent connected vehicles.
it is reported that the 2024 version of c-icap privacy protection has attack and defense protection evaluation indicators under the networked communication protection section, and is evaluated through two attack methods: real attack on the whole vehicle and simulated combined attack - checking whether there are security loopholes and defense weaknesses in the internet of vehicles system, and at the same time evaluating the vehicle system's defense capabilities against network malicious behavior, remote penetration and other attack methods.
multi-path and multi-scenario detection
application value and application security are the two wings of the development of end-to-end ai big models. at present, end-to-end ai big models have entered a period of rapid iteration. while actively exploring the implementation of applications, car companies must also attach great importance to security issues and maintain the bottom line of user privacy protection. china automotive evaluation has bravely taken on the important task of "national team of automotive evaluation". through c-icap privacy protection related evaluations, it helps the industry discover potential network attack defense vulnerabilities, effectively improves the level of product privacy protection, empowers the security of the internet of vehicles system, and promotes the healthy development of the intelligent connected vehicle industry.
