Ant Group released AI security detection platform "Ant Verification 2.0", which can detect the generative model with generative capability

Tech 2023-07-08 15:17:32 Source: Network

AIGC security has become an important issue in the current research and development of artificial intelligence. He Jifeng, Academician of the CAS Member, delivered a keynote speech titled "Credible AI Outlook"

AIGC security has become an important issue in the current research and development of artificial intelligence. He Jifeng, Academician of the CAS Member, delivered a keynote speech titled "Credible AI Outlook". He believes that "artificial intelligence can no longer be seen as a simple technical task, but as an engine of change for the entire organization. Trusted artificial intelligence should be committed to ensuring data security and credibility, system behavior accountability, algorithm model interpretability, network environment trustworthiness, legal and ethical trustworthiness, and other issues

At the forum, Ant Group also announced that "Ant Intelligence 2.0" is fully open, providing three detection tools for global developers, namely, AIGC security, AI interpretability and AI robustness, free of charge, which can serve large-scale complex business scenarios in digital finance, education, culture, medical care, e-commerce and other fields.

It is understood that "Ant Detection 2.0" has two hardcore capabilities, one of which is the ability to detect generative models using generative capabilities. The detection standards cover three major categories: content security, data security, and ethical security. It can perform risk confrontation detection on the generative content of the large model, including hundreds of dimensions such as personal privacy, ideology, illegal crime, prejudice, and discrimination. It will also generate detection reports to help the large model continuously optimize more targeted.

Li Junkui, president of the technology department of Ant Group's big security business group, said, Generative large models are a type of 'deep black box' technology. 'Ant Detection 2.0' uses intelligent game confrontation technology to simulate black production and automate the generation of massive test sets. It can achieve induced detection and calculation of 'generative robots' on' AIGC generative models', much like a 24-hour 'security hacker' searching for faults in the large model, in order to identify the weaknesses and security issues of the large model. This evaluation of AI by AI Teaching and learning are mutually beneficial. For example, generative examiners assess generative athletes, enhancing each other's abilities in confrontation, further enhancing the AI safety level of the entire industry

Another hardcore capability is the integration of interpretability detection tools into "Ant Detection 2.0". Combining AI technology and expert prior knowledge, through visualization, logical reasoning, Causal inference and other technologies, from 7 dimensions such as integrity, accuracy, stability and more than 20 evaluation indicators, the quantitative analysis of the interpretation quality of AI system helps users to more clearly verify and optimize the interpretable scheme.

It is understood that "Ant Detection 2.0" integrates ant trusted AI practices over the past 10 years, over 1000 trusted AI patents, and supports zero coding evaluation. Developers can identify and mine model vulnerabilities with just one click by integrating AI models into ant detection through API interfaces.

Ant Group believes that the deep application of AI not only puts forward higher requirements for data, algorithms and computing power, but also poses more challenges to security, privacy and ethics. Only by ensuring data security and privacy protection, as well as sound artificial intelligence ethics and security, can AI technology truly unleash its application value.

At this forum, AIGC Credible Initiative was jointly launched by China Academy of Information and Communication, Shanghai Artificial Intelligence Laboratory, Wuhan University, Ant Group and other units. This initiative has been participated by dozens of units. Focusing on the economic, security, privacy, data governance and other issues that AI may cause, it proposes to build a global governance cooperation framework for AIGC's credible development. It uses secure and credible data resources, technical frameworks, computing methods, and software platforms to comprehensively improve AIGC's trusted engineering capabilities, and maximize the security, transparency, and interpretation of generative AI.

The forum also released the White Paper on Trusted AI Technology and Application Progress (2023) jointly compiled by the Chinese Academy of Information Technology, Tsinghua University and Ant Group. This report is based on a unified cognitive dimension of trusted AI, summarizing the current development status of trusted AI, refining the new requirements triggered by the new stage of large models and AIGC, forming a technology based evaluation system and tools for trusted AI, and exploring the development of a new generation of trusted artificial intelligence.

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])

Mobile advertising space rental

Tag: Ant generative Group released AI security detection platform Verification