Openai will pay more than $ 500,000 for the discovery of errors and chatgpt failures

Openai has expanded the awards for those who find security vulnerabilities and errors in the operation of the company’s products and platforms. Artificial Intelligence Company (IA) is the owner of Chatgpt and is also a various licensed language models such as Whisper that transforms sounds into text and sora, a general video AI.

Last Wednesday (26) announced and emphasized a series of updates in internal cyber security programs The increase in the award of those who find critical violations in the services. The ceiling for an reward now $ 100,000 with a five -fold increase (or about $ 575,000 in direct currency transformation) Compared to the old value.

https://www.youtube.com/watch?v=RQBFIXQXEQW

This amount will be allocated “Important and High Effective Research on Safety“In particular, they help users directly. Also, Openai will give some payment bonuses to the senders of Openai for a limited period of time”Quality Reports“For the company.

Openai cyber security more news

The company is still concerned about the visible increase in Chatbot’s popularity and the expansion of its own services. Just as they reach more people and new products are released, platforms become more discontent.

Proof of this, An expansion of a cyber security institute financing and laboratories By Openai. In total, he Within two years, 28 initiatives in the region have financial assistance, but plans to expand access project. Idea is to focus on more specific areas that directly affect the brand, such as language privacy, permanent threat detection and autonomous protection such as the last operator.

Recently, a ChatGPT failure has been investigated in cyber attacks on financial institutions. In addition, 20 million chatbot users were offered for sale by a Russian invader who will provide some access to the company’s servers.

How do chatgpt errors work?

Call prize beetleOr Bug Hunt is already a traditional service in the digital security industry. Companies such as Microsoft, Meta and Google continue Reward Programs that benefit professionals, experts, independent researchers and enthusiasts who send complaints about security gaps In your products.

These partnerships are well seen by these companies and also the relevant researchers. Companies may correct violations that the internal team cannot detect for various reasons and may be a headache in the future, and that the field profession is rewarded by business.

In the case of Openai, the supply and control steps of this service are performed by the bugcrowd platform. In total, The company has already corrected 209 security vulnerabilities for contributions on the siteWhat performance did he perform Average $ 500 (or almost $ 2.9 thousand) payments in the last three months The complaint was accepted.

All information about Openai’s award -winning program can also be read in this link where you can register to begin to contribute. The financing scholarship for Cyber ​​Security Institutes can be requested on the company’s own page by completing a form in the proposed project.

Tecmundo spoke with Hacker, who said he had entered Oracle servers and stole data from more than 6 million people. Click here to check how the conversation went and get more information about the case.