Data is essential for both tools, but Data Protection is differentiated and differentiating
In the OpenAI Terms and Conditions, Article 3 states that the user will own the content generated by the tool. However, the user is also considered to be solely responsible for the content, either for the information transmitted by the user or for the content generated by the tool. If the content generated infringes the rights of third parties, the user will be held responsible, without the possibility of turning to OpenAI.
The risk of plagiarism is inevitable with the number of data ChatGPT uses from various sources
It is unlikely that openAI has obtained the permission of each author for the re-use of the content introduced for learning ChatGPT. ChatGPT also does not include its sources in its responses to users, which could potentially infringe on copyright.
In January, several companies filed a lawsuit against GitHub, OpenAI, and Microsoft for obtaining licensed source code to design the Copilot solution. These companies claimed that the tool could violate copyright if it was able to fully reproduce the code in a video game or other licensed software.
SYSTRAN guarantees :
- Data Protection: SYSTRAN uses technical measures to protect the translation data of its clients, especially sensitive and confidential data. They are not preserved or used for model training.
- Transparency: SYSTRAN provides detailed reports on translations performed for its clients, including originals and translations, to ensure transparency and traceability of translation processes.
- Data Ownership: SYSTRAN guarantees that all translation data of its clients is the exclusive property of its clients, and that SYSTRAN has no rights to the data. Systran does not store translated data, nor does it use it to train its models. Files, for example, are deleted after three days.
- Protection of intellectual property rights: SYSTRAN is committed to respect the intellectual property rights of its clients, in particular the copyrights and trademarks and also the intellectual property rights of the corpus of data used for translation models.
- Monitoring: SYSTRAN continuously monitors the activities of its employees and partners to ensure that its clients’ translation data is protected from unauthorized use.
Cyberhaven has published a study in which it is indicated that 11% of the data employees post on ChatGPT is confidential.
More and more businesses, public institutions and teaching are restricted or prohibited use of ChatGPT.
SAMSUNG, JP MORGAN, VERIZON … and AI AND MICROSOFT
Samsung group had authorized engineers from the Samsung Semiconductor branch, responsible for designing electronic components such as memory, storage, processors or photo sensors, to use ChatGPT to correct source code problems.
Samsung knowingly revealed the source code for a new program that is now stored on OpenAI servers
Samsung has decided to restrict the use of Chat GPT, already in force at JP Morgan and Verizon.
In addition, SYSTRAN has an on-premise solution:
SYSTRAN Pure Neural® Server is the non-premise solution that can be deployed behind the corporate firewall or hosted by SYSTRAN in a secure private cloud. The installation is tailored to the standards of enterprise performance and availability
This secure architecture provides guaranteed confidentiality and security and is in accordance with the internal data security policy: customer data, confidential and critical information do not leave the corporate information system.