Danielle CampbellSenior Project Manager| PM Consulting GroupSan Diego, United States
If you are using AI for specific trends, how do you protect data? Saving Changes...
Sort By:
Tola KehindeConsultant| Cedrusville CorporationWoodbridge, Ontario, Canada
From my perspective as an AI Consultant, the best practice is to mask PII elements of the data without stripping the relevancy and context of the data. Numerous off-the-shelf tools can do this.With tools like GPT, another option is to adjust the privacy settings such that your data remains proprietory to you and will not be used to further train the model.
...
3 replies by Allen Watkins, Danielle Campbell, and Rami Kaibni
Sep 04, 2024 7:37 AM
Rami Kaibni
...
Tola, interesting feedback. I didn't know we can adjust privacy settings in GPT such that your data remains proprietary to you and not be used to further train the model.
Sep 04, 2024 1:23 PM
Danielle Campbell
...
Thank you Tola, this is very helpful!
Sep 05, 2024 9:48 AM
Allen Watkins
...
" Numerous off-the-shelf tools can do this." Which tools have you used and are there any that you would say are better than others relative to ease of use and or technical ability required to properly manipulate the system to achieve the results you are seeking?
data Encryption and decryptions implementations helps ensuring data is securely encrypted both in transit and at rest. This reduces the risk of data being accessed if intercepted.
...
1 reply by Tola Kehinde
Sep 06, 2024 11:42 AM
Tola Kehinde
...
Yes, but encrypted data are no good for LLM (i.e. ChatGPT) to transform as they are unreadable until decrypted. I believe masking data's PII element is the way to go.
Senior Projects Manager | Field & Marten AssociatesNew Westminster, British Columbia, Canada
Sep 03, 2024 3:53 PM
Replying to Tola Kehinde
...
From my perspective as an AI Consultant, the best practice is to mask PII elements of the data without stripping the relevancy and context of the data. Numerous off-the-shelf tools can do this.With tools like GPT, another option is to adjust the privacy settings such that your data remains proprietory to you and will not be used to further train the model.
Tola, interesting feedback. I didn't know we can adjust privacy settings in GPT such that your data remains proprietary to you and not be used to further train the model.
...
1 reply by Tola Kehinde
Sep 06, 2024 12:09 PM
Tola Kehinde
...
Oh yes. There are lots of hacks with GPT, and it evolves constantly. I am launching a Masterclass in October for Professionals and Business Owners, where I will train and unveil leveraging numerous AI tools for Productivity and Profitability.
Saving Changes...
Danielle CampbellSenior Project Manager| PM Consulting GroupSan Diego, United States
Sep 03, 2024 3:53 PM
Replying to Tola Kehinde
...
From my perspective as an AI Consultant, the best practice is to mask PII elements of the data without stripping the relevancy and context of the data. Numerous off-the-shelf tools can do this.With tools like GPT, another option is to adjust the privacy settings such that your data remains proprietory to you and will not be used to further train the model.
Thank you Tola, this is very helpful! Saving Changes...
Allen WatkinsIT Infrastructure Manager| Greenville Utilities CommissionStantonsburg, Nc, United States
Sep 03, 2024 3:53 PM
Replying to Tola Kehinde
...
From my perspective as an AI Consultant, the best practice is to mask PII elements of the data without stripping the relevancy and context of the data. Numerous off-the-shelf tools can do this.With tools like GPT, another option is to adjust the privacy settings such that your data remains proprietory to you and will not be used to further train the model.
" Numerous off-the-shelf tools can do this." Which tools have you used and are there any that you would say are better than others relative to ease of use and or technical ability required to properly manipulate the system to achieve the results you are seeking?
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
AI is a board term. Protect data when some AI component is in use is the same than other type of initiatives where is not used. But changes in this matter appeared when the new model for generative AI was created (the basement for things like ChatGPT). In this case, the "scope" has been extended because it not only about protect data is about to Responsible AI. You can find a lot about that outside there but the three courses published by the PMI, mainly the 3rd, are a good source of information Saving Changes...
Tola KehindeConsultant| Cedrusville CorporationWoodbridge, Ontario, Canada
Sep 04, 2024 12:22 AM
Replying to AGADA CHARLES
...
data Encryption and decryptions implementations helps ensuring data is securely encrypted both in transit and at rest. This reduces the risk of data being accessed if intercepted.
Yes, but encrypted data are no good for LLM (i.e. ChatGPT) to transform as they are unreadable until decrypted. I believe masking data's PII element is the way to go. Saving Changes...
Tola KehindeConsultant| Cedrusville CorporationWoodbridge, Ontario, Canada
Sep 05, 2024 9:48 AM
Replying to Allen Watkins
...
" Numerous off-the-shelf tools can do this." Which tools have you used and are there any that you would say are better than others relative to ease of use and or technical ability required to properly manipulate the system to achieve the results you are seeking?
Tola KehindeConsultant| Cedrusville CorporationWoodbridge, Ontario, Canada
Sep 04, 2024 7:37 AM
Replying to Rami Kaibni
...
Tola, interesting feedback. I didn't know we can adjust privacy settings in GPT such that your data remains proprietary to you and not be used to further train the model.
Oh yes. There are lots of hacks with GPT, and it evolves constantly. I am launching a Masterclass in October for Professionals and Business Owners, where I will train and unveil leveraging numerous AI tools for Productivity and Profitability. Saving Changes...
Community Champion
