I've written before about the need to fact-check answers from general purpose AI tools such as...
Businesses Right to View AI Cautiously
First, let me state that I am slowly coming around on AI. I have found more useful purposes, such as analyzing chunks of data, summarizing survey comments, and creating buyer personas.
I also remain cautious about its use. Apparently, I am not alone.
While the use of Artificial Intelligence by businesses in the U.S. lags far behind the hype, the latest results of the U.S. Census Bureau's Business Trends and Outlook Survey do indicate that more businesses will consider implementing AI the remainder of this year.
As of February, 2024 ...
- Only 5.4 percent of U.S. firms used AI in producing goods and services in the previous two weeks (but up from 3.7% in September 2023)
- 84.2 percent did not
Over the next six months, however ...
- 6.6 percent of businesses definitely plan to use AI
- Those not planning to drop to 70.5 percent
- The "Don't Knows" doubled from 10.8 percent to 22.9 percent for the next six months, signaling that they plan to at least consider implementing AI by the end of 2024.
Why have American businesses been slow to put AI into production?
First, the Census Bureau acknowledges that the numbers could be low because businesses don't understand how much they use AI.
"There are many challenges in measuring AI business use that need to be recognized in interpreting our findings," the Bureau notes. "Firms may use AI in incidentally or insignificant ways, and not report such use, as AI may not be a part of their core processes or products. Others may rely on AI embedded in systems or services provided by third parties; in which case they are unknowing
users of AI. Still, we think that the BTOS statistics are highly relevant as they arguably reflect more direct and significant use of AI for business purposes."
So, while the use of AI by businesses is likely higher than the BTOS figures suggest, in another survey by Accenture, 72 percent of C-suite executives confirmed their cautious approach to AI investments because of social pressures to use AI ethically.
Businesses have taken note of news stories, Congressional hearings, privacy and consumer group concerns, and the shortcomings AI creators themselves acknowledge.
Anecdotal tales of issues with AI -- misrepresentations, stereotyping, discrimination, ethical, security, transparency, accountability, privacy, outdated knowledge, lack of sourcing, and occasional "hallucinations" -- have made businesses hold back too.
In a webinar for the Association of Independent Information Professionals, data scientist and cognitive research scientist Daniel Barulli spoke about the best practices of AI usage.
He mentioned implementing ethical guidelines, using "humans in the loop" of AI workflows to check AI's work, checking for plagiarism, ensuring that every citation in a document actually exists, and trusting only the consensus output of multiple AI tools; not just the output of any single one.
That's a lot for businesses to consider when thinking about implementing AI in a production capacity.
AI will get better as its training expands and improves. For now, perform your due diligence by being the human in the loop of your company's AI ventures.
You can also request an online meeting.