AI Commitments: But What about Chipmunks and the Bunny Rabbits?

July 23, 2023

Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.

Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI sent executives to a meeting held in “the White House” to agree on some ground rules for “artificial intelligence.” AI is available from a number of companies and as free downloads as open source. Rumors have reached me suggesting that active research and development are underway in government agencies, universities, and companies located in a number of countries other than the U.S. Some believe the U.S. is the Zoe of AI, assisted by Naiads. Okay, but you know those Greek gods can be unpredictable.

Thus, what’s a commitment? I am not sure what the word means today. I asked You.com, a smart search system to define the term for me. The system dutifully return this explanation:

commitment is defined as “an agreement or pledge to do something in the future; the state or an instance of being obligated or emotionally impelled; the act of committing, especially the act of committing a crime.” In general, commitment refers to a promise or pledge to do something, often with a strong sense of dedication or obligation. It can also refer to a state of being emotionally invested in something or someone, or to the act of carrying out a particular action or decision.

Several words and phrases jumped out at me; namely, “do something in the future.” What does “do” mean? What is “the future?” Next week, next month, a decade from a specific point in time, etc.? “Obligated” is an intriguing word. What compels the obligation? A threat, a sense of duty, and understanding of a shared ethical fabric? “Promise” evokes a young person’s statement to a parent when caught drinking daddy’s beer; for example, “Mom, I promise I won’t do that again.” The “emotional” investment is an angle that reminds me that 40 to 50 percent of first marriages end in divorce. Commitments — even when bound by social values — are flimsy things for some. Would I fly on a commercial airline whose crash rate was 40 to 50 percent? Would you?

“Okay, we broke the window? Now what do we do?” asks the leader of the pack. “Run,” says the brightest of the group. “If we are caught, we just say, “Okay, we will fix it.” “Will we?” asks the smallest of the gang. “Of course not,” replies the leader. Thanks MidJourney, you create original kid images well.

Why make any noise about commitment?

I read “How Do the White House’s A.I. Commitments Stack Up?” The write up is a personal opinion about an agreement between “the White House” and the big US players in artificial intelligence. The focus was understandable because those in attendance are wrapped in the red, white, and blue; presumably pay taxes; and want to do what’s right, save the rain forest, and be green.

Some of the companies participating in the meeting have testified before Congress. I recall at least one of the firms’ senior managers say, “Senator, thank you for that question. I don’t know the answer. I will have my team provide that information to you…” My hunch is that a few of the companies in attendance at the White House meeting could use the phrase or a similar one at some point in the “future.”

The table below lists most of the commitments to which the AI leaders showed some receptivity. The table presents the commitments in the left hand column and the right hand column offers some hypothesized reactions from a nation state quite opposed to the United States, the US dollar, the hegemony of US technology, baseball, apple pie, etc.

Commitments	Gamed Responses
Security testing before release	Based on historical security activities, not to worry
Sharing AI information	Let’s order pizza and plan a front company based in Walnut Creek
Protect IP about models	Let’s canvas our AI coders and pick some to get jobs at these outfits
Permit pentesting	Yes, pentesting. Order some white hats with happy faces
Tell users when AI content is produced	Yes, let’s become registered users. Who has a cousin in Mountain View?
Report about use of the AI technologies	Make sure we are on the mailing list for these reports
Research AI social risks	Do we own a research firm? Can we buy the research firm assisting these US companies?
Use AI to fix up social ills	What is a social ill? Call the general, please, and ask.

The PR angle is obvious. I wonder if commitments will work. The firms have one objective; that is, meet the expectations of their stakeholders. In order to do that, the firms must operate from the baseline of self-interest.

Net net: A plot of techno-land now have a few big outfits working and thinking hard how to buy up the best plots. What about zoning, government regulations, and doing good things for small animals and wild flowers? Yeah. No problem.

Stephen E Arnold, July 23, 2023

Written by Stephen E. Arnold · Filed Under AI, Business strategy, Government, News

Comments

Comments are closed.

Search the site
Subscribe to Beyond Search
Feature archive
News archive

Stephen E. Arnold monitors search, content processing, text mining and related topics from his high-tech nerve center in rural Kentucky. He tries to winnow the goose feathers from the giblets. He works with colleagues worldwide to make this Web log useful to those who want to go "beyond search". Contact him at sa [at] arnoldit.com. His Web site with additional information about search is arnoldit.com.