Jean Laval Chue HimDirector| Stella Aurorae Accountants Pty LtdSydney, Nsw, Australia
AI Explainability: Do you agree with the new laws that the results of AI programs should be able to be traced from input to output just like for an Object Oriented Programming software, no matter how many layers of abstractions. Auditors should be able to map the AI steps and explain the LOGIC. AI should not be a Black Box. Otherwise, AI will not be an enabler in decision-making Saving Changes...
Senior Projects Manager | Field & Marten AssociatesNew Westminster, British Columbia, Canada
Jean, the situation you're explaining is an ideal one but I don't know at this point if this can be achievable. It's still a vague area! Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
This is not new. It is there from more than 50 years ago. The "problem" IA becomes a buzzword from the point in the time Google publish "Attention is all you need" paper and OpenAI generated chatgpt based on that making it public and free of use. Before that, all the work on Ai was more in the field of software engineering. For example, human being are surrounded of Ai in things like refrigerators and air conditioning which from more than 30 years ago bring fuzzy logic into them. Just to comment, because I am fully involved into it, the use of generative Ai at organizational level demands the creation of almost a new business unit where you need to hire lawyers, linguistic, inclusion specialist adding to technicall people obviously.
...
1 reply by Jean Laval Chue Him
Apr 19, 2024 9:12 PM
Jean Laval Chue Him
...
Hi Sergio
thank you for your comments. However, I think that if we need a new business unit that includes Lawyers, this means that these AI systems results cannot be validated for critical use. And, we will need specialist lawyers to defend how the management made its decisions based on these AI systems and say environmental issues had not been taken into consideration or other life experience factors have not been taken into consideration by the AI systems
While you'd certainly like to have traceability, I believe the expectation of explaining HOW the results were guessed is unrealistic for any moderate to high complexity learning models, especially those that have been able to derive hitherto undetermined patterns from their training data.
This is why practices such as validation testing is in the pharma industry might need to evolve to apply to AI models.
Kiron Saving Changes...
Jean Laval Chue HimDirector| Stella Aurorae Accountants Pty LtdSydney, Nsw, Australia
Hi All,
thanks for your comments. Much appreciated. I am talking in the view of an AI Auditor. To make decisions based on results given by AI one should be able to explain logically HOW Logically it arrived at these advices. Only then AI can be adopted in critical use areas domains Saving Changes...
Jean Laval Chue HimDirector| Stella Aurorae Accountants Pty LtdSydney, Nsw, Australia
Apr 19, 2024 5:43 AM
Replying to Sergio Luis Conte
...
This is not new. It is there from more than 50 years ago. The "problem" IA becomes a buzzword from the point in the time Google publish "Attention is all you need" paper and OpenAI generated chatgpt based on that making it public and free of use. Before that, all the work on Ai was more in the field of software engineering. For example, human being are surrounded of Ai in things like refrigerators and air conditioning which from more than 30 years ago bring fuzzy logic into them. Just to comment, because I am fully involved into it, the use of generative Ai at organizational level demands the creation of almost a new business unit where you need to hire lawyers, linguistic, inclusion specialist adding to technicall people obviously.
Hi Sergio
thank you for your comments. However, I think that if we need a new business unit that includes Lawyers, this means that these AI systems results cannot be validated for critical use. And, we will need specialist lawyers to defend how the management made its decisions based on these AI systems and say environmental issues had not been taken into consideration or other life experience factors have not been taken into consideration by the AI systems
...
1 reply by Sergio Luis Conte
Apr 20, 2024 6:16 AM
Sergio Luis Conte
...
What I stated related to create a new business unit is regarding the use of generative AI. Regards.
Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
Apr 19, 2024 9:12 PM
Replying to Jean Laval Chue Him
...
Hi Sergio
thank you for your comments. However, I think that if we need a new business unit that includes Lawyers, this means that these AI systems results cannot be validated for critical use. And, we will need specialist lawyers to defend how the management made its decisions based on these AI systems and say environmental issues had not been taken into consideration or other life experience factors have not been taken into consideration by the AI systems
What I stated related to create a new business unit is regarding the use of generative AI. Regards.
...
1 reply by Jean Laval Chue Him
Apr 22, 2024 7:29 PM
Jean Laval Chue Him
...
understand your point. Thanks.
Saving Changes...
Jean Laval Chue HimDirector| Stella Aurorae Accountants Pty LtdSydney, Nsw, Australia
Apr 20, 2024 6:16 AM
Replying to Sergio Luis Conte
...
What I stated related to create a new business unit is regarding the use of generative AI. Regards.
understand your point. Thanks.
...
1 reply by Sergio Luis Conte
Apr 23, 2024 5:15 AM
Sergio Luis Conte
...
you are welcome. AI is a boarder term that, in my personal opinion, is becoming a new buzzword because generative AI is making a great job on publicity. But AI is using from more than 40 years ago in lot of type of initiatives. For example, in manufacturing or in personal cell phones.
Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
Apr 22, 2024 7:29 PM
Replying to Jean Laval Chue Him
...
understand your point. Thanks.
you are welcome. AI is a boarder term that, in my personal opinion, is becoming a new buzzword because generative AI is making a great job on publicity. But AI is using from more than 40 years ago in lot of type of initiatives. For example, in manufacturing or in personal cell phones. Saving Changes...
Community Champion
