AI and humans
Getty
I recently attended and moderated a session at the Reuters Thompson Momentum Conference in Austin, Texas. My session was “Hype and Fear: Managing Consumer Perceptions and AI Experiences.” The conference covered many topics such as trends in AI investment and employment, how AI works today, how people use AI in their daily work, and of course what to worry about about AI.
Joining my panel discussion session were Sandeep Dave, Chief Digital & Technology Officer at commercial real estate company CBRE, Lauren Kuntze, CEO of AI company ICONIQ AI, Sri Sivananda, Executive Vice President and CTO of PayPal, and Dr. Jesse Ehrenfeld, President of the American Medical Association. I chaired the session on behalf of IEEE.
At the outset, I pointed out that many people at IEEE are working on AI and its applications, and that IEEE publishes many conferences and publications that focus on the technical and practical application of AI. In addition, the IEEE Standards Association (SA) has organized several activities on AI ethics. In 2016, IEEE SA released the first edition of its report on Ethically Consistent Design of Autonomous and Intelligent Systems. The latest update to this report was released in 2019. Below is a history of IEEE activity on the use of AI.
ieEE standard activities addressing ethics and AI
IEEE SA Briefing
IEEE has worked with policy makers around the world to outline the appropriate uses of AI. Several IEEE standards have been written or are forming to address the ethical use of AI. These include standards on many topics such as data privacy, algorithmic bias, child and student data governance, transparency of autonomous systems, and more. The IEEE SA is also considering working with third parties on certifying AI practices that comply with evolving AI standards.
Here are some comments from other panel members. Sandeep Dave said AI/ML is not new and they are already implementing ML in several areas of the real estate lifecycle to improve efficiency, predict (market movements, asset failures, etc.) and predict. We believe there are significant AI benefits across the lifecycle of real estate, from doing things differently (i.e., significant productivity gains) to doing things differently (e.g., iterative generative design towards desired goals, combining GenAI and visualization capabilities to “experience” spaces prior to construction).
Lauren Kunze said generative AI is overhyped in the sense that companies still need rules. Many companies will make costly mistakes because they don’t understand what she can and can’t do with AI. She also said AI will disrupt the way we work and the jobs of the future. As an example, she cited the case of women’s fashion company H&M working with her Meta to promote authentic fashion products for her company using her ICONIQ AI virtual girlfriend creator Kuki (@kuki_ai) on Instagram.
Sri Sivananda said that fears, anxieties and doubts are real and still exist. He said AI is an enabler and a means, but not an outcome in itself. Using AI to optimize results requires trust, and he believes true innovation comes from two opposing challenges: first, harnessing its power, and second, making it human.
Dr. Jesse Ehrenfeld said the AMA House of Representatives will develop principles and recommendations regarding the benefits and unintended consequences of AI-generated medical advice. He said doctors are embracing emerging technologies but cannot ignore reliability, regulatory and public policy concerns. He also said that cutting-edge AI-enabled tools are still unable to diagnose and treat disease.
There are other considerations to keep in mind when using modern, complex AI models. Sandeep Dave said there is a real cost to deployment and enterprise access can quickly become expensive and ROI difficult. Furthermore, he said the use of GenAI may go against the organization’s sustainability goals due to the energy expended in training these models.
One of the topics discussed during the panel discussion was trust in AI when it is not clear how it operates. Many AI algorithms develop complex weighting models based on recognizing patterns in data. These models do not follow normal human reasoning methods, which can lead to skepticism about what the models are doing and how useful the end results are.
Clearly, knowing more about what AI is doing and how it makes decisions can go a long way in understanding what it can and cannot do, which in turn helps us use it in an ethical manner.
follow me twitter Or LinkedIn. check out my website.
