Work in Progress

by Katherine AUGUST, PhD., Biomedical Comms. Engineer
“AI for Human Rights, Emotional Integrity, and Public Trust“
Presented at the LMGlobal.Org‘s panel on “Towards a Public Discussion for Shaping the Future of A.I. Regulations in the U.S. & Beyond“, which virtually presented in collaboration with Bons-Ai.Org on November 15, 2025.
Presentation (PDF)

This discussion explores how engineers, policy makers, stakeholders, and both local and global communities can work together to advance human-centered, responsible AI.

[LMGlobal.Org] First, I want to say these are my own views. I am not speaking on behalf of any organization. Many people have contributed to this work, and I want to acknowledge them.
I’ve been very involved with a volunteer engineering organization of about 500,000 members globally, divided into 39 different societies. One of them is the standards organization. I’ve become very involved with standards. Thank you, Professor Victor Lawrence, for getting me involved with IEEE and standards, and thanks to everyone here who knows or works with these efforts.

This talk explores how engineers, policymakers, stakeholders, and both local and global communities can work together to advance human-centered, responsible AI. One of the standards groups I’ve worked with recently focuses on the social implications of technology. My specific working group has been called ‘Transparent Design for Wellbeing.’

We actually had to change that name and parts of our work. Our current president has issued mandates about certain words we cannot use. So we have had to edit many of our concepts, documents, and vocabulary. That is personally challenging, and I am very sensitive to it. If anyone has suggestions, I would welcome them.

This slide shows just a few of the people involved. From time to time, we have had to change what we call ourselves. You can see one of the older names, ‘Accessible and Responsible Technology.’ We are, in a sense, trying to fly under the radar.

We believe people everywhere are navigating a rapidly evolving AI environment, often without the guardrails they might expect from systems that move at human pace rather than digital speed. This affects education, workplaces, social interactions, personal devices, transportation, banking, and healthcare in many unexpected ways.

I appreciate the earlier mention of storytelling. Some of these images tell stories: perhaps a robot that might help us in old age, maybe even folding our laundry. We’re not sure how that will play ou
We also have navigation and traffic applications. Some may remember ‘Snowmageddon’ in western Manhattan a few years ago, when navigation apps directed huge numbers of drivers into the same place during a snowstorm. Even people who knew where they were going followed these apps almost mindlessly. There are many similar examples in daily life. People can lose skills they once had and become dependent on automated guidance.

Technology can also ease pressures, especially in learning. Some teachers on this call know that learning core skills takes time. It is often faster to write a program or use AI to generate code than to learn the underlying concepts. People might get a good job this way, but are they being fair to themselves now and in the future if they skip foundational understanding?
We need to ask: how do we support people through the whole lifespan so they can live well and stay “fit” in this world, not only physically but mentally and emotionally?

Presentation (PDF)

Another issue is social and emotional pressure. People experience bullying, toxic interactions, and overwhelming exposure online and offline. Could we use technology to monitor heart rate, breathing, and skin response to signal when someone has had too much exposure to toxic environments? Could parents, teachers, caregivers, and others adjust tasks or environments to protect people’s emotional state? We might be able to design guardrails that allow us to optimize daily experience and emotional well-being.

At the same time, many devices now lose or change functionality in unexpected ways as software updates or services change. We need to help people cope and adapt to this constant flux of technology in their homes and lives.

Earlier, someone spoke about AI in medical care. IEEE has a flagship magazine, IEEE Spectrum, sent to members across all 39 societies. It often highlights key technologies. I recently tried to emphasize how algorithms now process decisions and data at speeds beyond human perception, creating a gap that demands oversight, transparency, and public trust.

Even below the top-level models, we must understand how underlying data and physics interact. In fields such as radiology and medical imaging, the way data are captured and processed can create combinations that are not compatible or safe. When people talk about FDA oversight and “software as a medical device,” new questions arise that existing processes may not fully answer. We need to rethink approval methods and how we evaluate safety and performance.
When a clinician looks at an image, they want to trust it. Sometimes they can, sometimes they cannot. This pattern is repeating in many sectors.

Another IEEE activity is Congressional Visits Day in the United States. Once a year, IEEE members are invited to travel to Washington, D.C. at their own expense. We are not funded for this; we volunteer our time and pay our own way. We visit members of Congress, bring forward technology issues, and serve as technical consultants. This is important so lawmakers understand how technology works and can make better decisions.

However, while other nations are accelerating, we recently had a U.S. government shutdown. Many things stopped. The pipeline of new technology and regulation depends on a functioning government and functioning teams across society. Education, research investments, grants, universities, agencies—all of that depends on a stable system.

Because of that pause, I believe we are not just months behind; we are probably years behind. That lag is not easy to overcome. It is not only a loss of legislation. We lose momentum in intellectual property, research, and global leadership at a time when AI is advancing rapidly elsewhere.
We need progress not just for technology’s sake, but for safety, security, and well-being of individuals, industry, and society.

Engineers also look at how intellectual property law interacts with AI. Patents and copyright help people compete, prosper, and leave something meaningful behind. Among the top 10 AI patent holders in the world, only three are U.S. organizations: IBM, Alphabet, and Microsoft. Microsoft, importantly, is deeply involved in document management.

In many industries—including medicine—doctors now write their notes in AI-enabled document systems. These tools can prompt the doctor to record a better, more complete account of what happened, and to show that they followed “standard of care,” because the form itself “knows” what standard of care is.
This can be good, but it also introduces bias toward the standard of care, which is based on a limited set of prior cases that may or may not reflect the physiology or situation of the patient in front of them. These are the kinds of “wildcard” use cases we worry about. We need ways to capture those stories, process the information, and iteratively adjust, because these patterns will propagate through critical systems in ways we do not fully understand.

We see this also in search and retrieval. Many questions we pose to AI systems still have unexplained or incorrect outcomes. I have recently searched for information I absolutely know exists—such as certain patents or documents I worked on—in Google or ChatGPT and received wrong answers or could not find them at all. We are told that everything ever written is accessible to these engines, yet some known items do not appear.

We must test these systems repeatedly and systematically. Some problems do not require searching the entire internet. They involve finite, stable datasets. Those cases may be better served by older methods such as hidden Markov models or other contained approaches, rather than large, general models.
IEEE also produces a monthly video update on U.S. legislation with Eric Heilman, an IEEE staff member. He summarizes what is happening in Washington with respect to technology, including AI. It is very helpful for engineers and citizens.

Recently, he has discussed the AI LEAD Act, the AI Risk Evaluation Act, and the AI for America roadmap, associated with Senator Mark Kelly and others. These are familiar names in the news, but it is helpful to have focused summaries. These videos are available online.

He notes that AI remains one of IEEE-USA’s strongest and most active policy areas, with ongoing engagement on Capitol Hill. Engineers are often asked to advise on technical aspects of bills and ways to improve them.

In November’s update, there were also notes on work to protect children from AI chatbot harm, which connects back to some of the standards topics I mentioned earlier.

One of our new standards efforts focuses on transparency, accuracy, and source integrity in media and public-facing information. It is an emerging framework to strengthen AI systems by improving trust, informational integrity, and emotional integrity—understanding how repeated exposure to certain content affects people’s emotional state and well-being.

We want technology, including 5G and future 6G networks, to support these goals. Because 6G is not yet fully defined, there is a real opportunity to embed these standards early and to shape how future systems support human values.

So, we can remain hopeful. There could be a nationally coordinated approach to AI regulation in the United States that reflects global best practices, with standards like our transparent design work contributing both internationally and domestically.

However, participation matters. Out of hundreds of thousands of IEEE members globally, only about 300 people went to Capitol Hill for Congressional Visits Day. Many more could engage with legislators and help them understand the issues

We should aim to uplift humanity. AI should help us do that. Thoughtful regulation and legislation can be more effective if technical experts and communities are active in the process.

We should aim to uplift humanity. AI should help us do that. Thoughtful regulation and legislation can be more effective if technical experts and communities are active in the process.

Across the country, states and federal lawmakers are exploring laws around transparency in AI systems, protections for minors and vulnerable populations, and workforce preparation for an AI-augmented economy. Every major technology transition affects the workforce. AI can analyze what people do on the job, learn how tasks are performed, and support performance. It can also help identify new forms of work that might make people happier.

Jobs will change. Some tasks will shift to technology, but that creates new opportunities. As a previous speaker said, those who understand their freedoms and possibilities can design new roles, new businesses, and new ways to participate—especially if they have access to open-source tools and learn to use AI from a young age.

We need to understand where “weird” data comes from, how to find good data, how to test outputs, and how to identify reliable results.

That is my message: IEEE is one of the few organizations positioned to bridge these worlds—technical, ethical, societal, and global—in the AI era.

Thank you.

Presentation (PDF)