Well Being: Human Centric Artificial Intelligence

One of the things we humans like is stability: stability in our society, stability in our daily lives, stability even in the prices of things like, you know, eggs.

Change, of course, is a threat to stability. For example, the inflation that most of the world has experienced in recent years. The simple change, the increase in the price of things, has not only been responsible for the increase in prices, but it has also had a side effect: the change of democratic governments around the world. The UK. The US. And soon Canada.

We always have some level of inflation, for example I remember comic books were $1.25 CDN and now are >$6, which seems pretty expensive, but that is slow, "normal" inflation.

It's the fast inflation, the high inflation of 5, 6, 7%, it's side effect is that it damages democracy. Too much, too fast, and with side effects that are difficult to understand.

Technology and Productivity

Alongside high inflation, technology is something that brings significant societal instability. Let's not forget the printing press, the loom, antibiotics, etc., but in the recent past, the last few decades, we've had innovations like computers, the capital "I" Internet, the cloud, mobile phones, and now the recent advances in AI transformer technology that enable generative AI.

While we deal with all this technical change, internally many of us expect that this is somehow making us better, faster, stronger, and presumably more productive. However, that is not always the case.

Despite the advent of personal computers, the internet and other high-tech innovations, much of the industrialized world is stuck in an economic growth slump, with O.E.C.D. countries expected to expand on aggregate just 1.7 percent this year. Economists sometimes call this phenomenon the productivity paradox. - NYTimes

ASIDE: Canada's productivity is of particular concern (perhaps "concern" isn't strong enough a word) and is worth understanding by anyone in any country, because the place where you live could have the same problem soon, if not now.

Unfortunately, since the pandemic, with the exception of agriculture, forestry, hunting and fishing, productivity growth within goods-producing industries has not only slowed but has outright reversed (Chart 2). From an annual increase of 1.4% in the decade prior to the pandemic, it has declined by 1.2% annually since.  - https://economics.td.com/ca-productivity-bad-to-worse#:~:text=Unfortunately%2C%20since%20the%20pandemic%2C%20with,declined%20by%201.2%25%20annually%20since.

This needs to change. We need more productivity.

Side Effects

While we are also seeing direct and significant disruption in society from some of the technologies we are building in terms of what we expect them to do, they also have initially unseen side-effects. For example, we expected social networking to allow people to connect through computers and algorithms, which I suppose it does, but we didn't expect, at least at the start, the negative side effects.

A social network...sounds perfectly "human-centered" but ends up being the opposite. We have managed to take a technology that on the surface is meant to connect people and do the exact opposite.

1. Polarization and exposure to misinformation
2. Pressure and anxiety
3. Disconnection from traditional media
4. Effects on voting behavior
5. Disinformation campaigns by foreign powers
6. Erosion of trust
7. Moderation of political positions
8. Body image issues
9. Cyber-bullying

The list goes on. The side-effects are considerable, and I think if we could turn back the clock we might make different technology, implementation, and educational choices.

Human Centered AI and Well-Being

With all this in mind, what will we do with Generative Artificial Intelligence (GenAI), a new and disruptive technology? What changes will it bring? What side effects will it have? How will we interact with it? Are we thinking big enough, or are we just doing the obvious implementations, much like we continue to build computers with a mouse, keyboard, and screen? Change is disruptive, perhaps more disruptive than we understand, for example it can even bring down governments, especially if it has a negative economic impact.

We need to think about what technology has done before and be a little more careful about how we apply GenAI, perhaps by applying Human Centered Artificial Intelligence (HCAI).

Human-Centered AI (HCAI) is an emerging discipline intent on creating AI systems that amplify and augment rather than displace human abilities. HCAI seeks to preserve human control in a way that ensures artificial intelligence meets our needs while also operating transparently, delivering equitable outcomes, and respecting privacy. - https://research.ibm.com/blog/what-is-human-centered-ai

...human-centered AI (HCAI) seeks to re-position humans at the center of the AI lifecycle (Bond et al., Citation2019; Riedl, Citation2019; Xu, Citation2019) and improve human performance in ways that are reliable, safe, and trustworthy by augmenting rather than replacing human capabilities (Shneiderman, Citation2020a). - https://www.tandfonline.com/doi/full/10.1080/10447318.2022.2153320#d1e645

We build technology to help us, to get productivity to improve our quality of life. Computers work for us, not the other way around.

Augment Humans, Don't Simply Replace Them

What we need to do is augment humans with AI, not replace them. If we can augment people to help us do our jobs faster and better, and maybe even give us more free time, that would actually make us more productive. That's what we're looking for, that's what we need to make our lives better AND make the stock market go up. We don't see continuous improvement in the stock market from disruptive technologies, we see it from continuous, long-term, incremental improvement in overall productivity.

Of the six challenges the paper identifies, one of them is human well-being, and it goes into what we need AI to do to make AI more human-centered and consider human well-being. Here are the paper's suggestions for considering human well-being in AI.

1. Study the potential benefits and harms of AI
2. Promote well-being in social media
3. Support and expand human cognitive capacities
4. Protect human attention and mindfulness
5. Adapt to humans
6. Respect human resources, including time and data
7. Support healthy human emotion management and social interaction
8. Adopt simplicity in design
9. Prevent negative side effects
10. Be inclusive
11. Avoid bias
12. Be transparent and accountable
13. Provide human controls, feedback, and affordances so users can calibrate their trust of AI outputs

Ultimately, implementing disruptive technology incorrectly costs us dearly in productivity. More to come!

Further Reading

Link - https://www.tandfonline.com/doi/full/10.1080/10447318.2022.2153320#abstract

Link - https://economics.mit.edu/sites/default/files/2024-05/The%20Simple%20Macroeconomics%20of%20AI.pdf

Link - https://www.nytimes.com/2024/07/13/business/dealbook/ai-productivity.html

Link - https://research.ibm.com/blog/what-is-human-centered-ai

Link - https://economics.td.com/ca-productivity-bad-to-worse#:~:text=Unfortunately%2C%20since%20the%20pandemic%2C%20with,declined%20by%201.2%25%20annually%20since.

Link - https://wellbeingairesearchinstitute.com/