The fake news problem is real. Today, there is no independent or trusted media left. Everyone has a bias and it has been amplified with social media echo chambers. The problem with governance is that there is no real way to be completely objective, since the bias of governing body is implied. If we could remove the element of human judgement from governance, it might be possible to have responsible AI
This week, there was a highly publicized incident where Google's AI model Gemini made an inaccurate claim that President Obama was the first Muslim president. This mistake underscores the critical importance of responsible AI development and trusted media. While AI offers incredible potential to augment human capabilities, incidents like this highlight how AI systems can also inadvertently perpetuate biases and spread misinformation if the source data is not accurate and the models are not developed with great care. AI models like Gemini essentially function as knowledge mines, piecing together plausible responses based on the data they are trained on. However, their training data and algorithms can reflect societal biases and false narratives present in the underlying information sources. This brings up a timely topic: Who should be responsible for filtering out fake news? Should we categorize the behavior of intentionally creating fake news and deep fake as illegal? Is it time to bring back the trusted & independent media practices we used to have? Feel free to share your thoughts on how we can collectively address the challenges posed by deepfakes and ensure the integrity of our data & AI landscape. #ResponsibleAI #Deepfakes #Misinformation #TrustedMedia #DigitalTrust #Innovation https://lnkd.in/gED_3dDq
Great point - Bias is when an insight or inference reflects partiality, preference, or prejudice for or against a person, entity, group, fact or idea. Question is how to determine it. Now bias can be due to Completeness of information gap Influence by selective choice of words or terms Inferenceabilty impact with missing information There are other people aspects to - are personalities of stakeholders influenced by certain drivers that can distort narratives. Of course , last but not least - is there a process in place which has been set objectively to detect bias that is well governed with well defined rules and policies. This is obviously something that needs some thinking to detect , monitor and mitigate/prevent bias.