How ChatGPT and SYSTRAN handle Ethics issues
Biases or discrimination
ChatGPT is driven on existing data, which can introduce bias into the model. If these bias is ignored, it can lead to discriminatory or unfair outcomes. CSA Research, modeled, in its last report, how Algorithmic Bias Has Many Causes:
ChatGPT collects information that can be false and delivers it without context.
Trusting ChatGPT entirely can be a big risk, if the information is harmful (for example, financial scam).
“The potential exploitation of these types of AI systems by criminals offers a bleak prospect,” the agency explains in its report. It also points to the possibility of the creation of dark LLMs, AIs created and trained solely for a criminal purpose and which could become central in a mafia business model. “It will become easier than ever for malicious actors to commit criminal activities without prior knowledge necessary,” the report notes.
SYSTRAN offers an ethical solution. Ethics is maintained throughout the translation workflow, on the corpus of incoming data and the results, each verified by linguists. Human intervention ensures respect for ethics without the algorithms multiplying potential errors.
Moreover, there can be no cyber crime problem, since both data and results are secure. Thus, SYSTRAN has gained the trust of the most demanding organizations: Defense, Security, Large Accounts, Public Organizations …
Compliance with RGPD
Two complaints were filed in France against Open AI and ChatGPT against the French Personal Data Congressman Cnil on Tuesday, April 4.
In her complaint reviewed by the AFP, Zoé Vilain, president of the digital advocacy association Janus International, explained that she created an account to use ChatGPT and found that there were no “Terms and Conditions” to accept and “any privacy policies”. She asks Cnil to help exercise her right to access her personal information collected by OpenAI, after an unsuccessful attempt at the company.
A second complaint was filed by David Libeau, a developer who is very committed to protecting personal data. He explains in his complaint that he has identified personal information about him by questioning ChatGPT on his profile. “When I asked for more information, the algorithm began to flourish and attributed to me the creation of websites or the organization of online demonstrations, which is completely wrong,” he wrote
On March 31, Italy was the first country to temporarily suspend ChatGPT, owing to concerns about the security of personal data, the lack of information provided to users, and the absence of an age filter to verify users. This measure has since attracted the attention of other European authorities, such as France, Ireland and Germany, who have contacted the Italian authorities in order to develop a common position towards ChatGPT.
SYSTRAN is RGPD compliant
- User Consent: SYSTRAN requires the consent of users to collect and process their personal data when using its translation services.
- Access Control: SYSTRAN ensures that only authorized employees and partners have access to translation data.
- User rights: SYSTRAN guarantees that users have the rights defined in the GPD, such as the right of access, the right of rectification, the right to erase and the right to portability of data.
- Data Transfer Security: SYSTRAN ensures that translation data transfers are secure, using appropriate encryption mechanisms when transmitting data.
- Data Retention – SYSTRAN retains translation data for the duration required to provide its services and in accordance with applicable legal requirements