The Question of AI: Latent Persuasion and Bias

“Autocratic governments could mandate that social media and productivity tools all nudge their citizens to communicate in a certain way.” Article by Christopher Mims, Wall Street Journal, May 2023, entitled: ‘Help! My Political Beliefs Were Altered by a Chatbot.’

? Have you already made up your mind about the value of AI? If so, it may be wise for you to step back once more and take time to reflect further on the pros and cons. AI is clearly an evolving, intriguing concept, so now’s the time to weigh-in before it becomes too established – for better or for worse. We need to shape it to co-opt as many benefits as possible, before some of its less attractive aspects potentially take hold, too.?

?Mims' sub-title read: ‘AI assistants may be able to change our views without realizing it.’ He quoted a professor from Cornell University who coined the term latent persuasion to define the phenomenon induced by any AI author’s intent of the moment. Adopting the hidden biases of its authors, one way or the other, is always a key possibility. Apparently our best defense against such subtle influences, as things stand today, is to become acutely aware of it. Open AI, sponsored by Microsoft, has been eager to encourage “responsible AI.”

Coupled with this, we have to be acutely aware of another human phenomenon, as pinpointed by Sir Joshua Reynolds an English philosopher of the eighteenth century: There’s no expedient than man will not go to avoid the labor of thinking! It was also attributed to Thomas Edison, too. If this is likely true, and there’s much anecdotal evidence to suggest that it is, then we should be ever more vigilant in our personal use of AI. Mims referred to products currently being produced by Google and Microsoft, which are nuanced to make us react more positively than we’re otherwise disposed to do.?

Another professor Mims quoted, pointed out “…that algorithms like ChatGPT don’t have beliefs,” but they “offer opinions” from the way they are trained and programmed and “those opinions can be measured.” Fortunately, at least one monitor of large-language scoops for AI purposes exists: The Holistic Evaluation of Language Models based out of Stanford. This covers taking into account the scrapes of wide internet swaths that AI authors scoop-up that invariably reflect opinions and biases inherent in the texts they absorb.?

This likely means that so much of AI output we currently absorb could be at the behest of “…clones of the perspectives and values of relatively young, college educated, West-coast-dwelling people.” In the event that is the case, we should be mindful of potential brainwashing from a variety of AI sources. AI could have a subtle influence over what we believe and the outcome of future elections: Another reason why behaviorists and politicians should examine creating AI principles to minimize the possibility of bias.

China has already taken a big lead in AI for political reasons among others and Russia is well invested in it, too. The recent walkout by Hollywood writers and producers was much to do with AI and its potential for replacing their role. Many fear that the use of AI could replicate actors or reproduce screenplays. Even so, there’s much evidence to suggest that we should welcome AI, even though we should do so with our eyes wide open and nudge our legislators hard to build in appropriate safeguards.

Perhaps it would be a good idea if we all do our best to start instigating a constructive way forward by:

? Deliberate over passing this article around within your circle and ask for people’s ideas on how things could be changed for the better

? Ponder asking friends and acquaintances if they know of people in influential positions, who would be interested in taking up this important torch

? Look at writing to your local newspaper editor and request that he/she give due importance to the topic

? Think about bringing the matter up at your next professional association, social club and/or social gathering sessions to encourage support for AI guidelines

? Consider writing to your Senator, Congress person, or prominent political/organizational leader and request they get serious about monitoring AI and encourage positive applications

? Contemplate sharing your perceived benefits and concerns about this issue at every worthwhile venue

Once public opinion gets behind the right movement, feel free to play your part in helping it to succeed.

Let’s get started!

________________________________________________________________

Author, Peter A. Arthur-Smith, Founding Principal with Leadership Solutions, Inc., is based in New York, and author of Smart Decisions: Goodbye Problems, Hello Options. He has drafted a potential new publication about enlightened leadership that offers a slew of fresh leadership concepts and practical models. Feel free to follow author at: Linkedin.com/in/peter-arthur-smith-2115722/?

? 2023 PAAS-LSI- All rights reserved.

?

.