Page 26 - April 2024 INSIGHT MAGAZINE.indd

BUSINESS TIPS

AI use in business requires

ethical balance

As technology continues to rapidly says, referring to the problems Microsot policy should also outline what workplace
evolve, businesses are increasingly experienced when it released its Tay bot devices can be used for AI. “If you need to

turning to Artificial Intelligence (AI) in March 2016. save something because you’re involved
to streamline operations, enhance The bot, under the name TayTweets in a lawsuit, then you don’t want to it

efficiency, and gain a competitive edge. with the handle @TayandYou, resulted be on an employee’s personal device
because you won’t have access to it.”
There is no question surrounding in Twitter (now known as ‘X’) users
the benefits of integrating AI into tweet politically incorrect phrases and As well, Professor Grossman also

business processes, but there remain inflammatory messages resulting in the recommends employees clearly know
legitimate concerns that accompany this bot releasing racist and sexually charged what AI tools are okay to use and which
technological leap. messages in response to other users. are not and ensure they are fully trained.
One primary concern is the ethical Initially, Microsot suspended the account “You don’t want them violating intellectual

implications of AI implementation. As ater 16 hours, erasing the inflammatory property rules or other privacy rights.

AI systems such as ChatGPT, ClickUp, tweets and two days later took it o¯ine. You also don’t want them putting into a

Copy.ai, or Kickresume become more “Most systems, like ChatGPT, are trained public tool any confidential or propriety
sophisticated, they oten require access on the internet and that has its pluses information,” she says. “Some companies

to vast amounts of data to function and minuses,” says Professor Grossman, have turned o† the ability to use these

e†ectively. This raises questions adding ‘hallucinations’ pose another big AI tools because they are terrified

about privacy and the responsible use problem for AI users. employees will put propriety information
of sensitive information, as well as out there while asking a question about
legal concerns surrounding the use of “ChatGPT for example is trained to a problem they are working on. If you’re
intellectual property. generate new content and to sound very using one of these open-source tools, it’s
conversational, so it uses what it has
“The question is fair use or is it a violation learned on the internet to predict the like Google or anything else; it’s free rein.”
of copyright,” says Maura Grossman, next most likely word. But that doesn’t Professor Grossman says rules and
Research Professor, School of Computer mean it’s telling you the truth.” regulations around AI will be gradually
Science at the University of Waterloo, strengthened, noting a new regulation
whose expertise centres on AI policy and She says there have been instances of coming into play in B.C. pertaining to
ethics. people using AI to conduct legal research issues surrounding intimate imagery is
and submitting bogus case citations in
She notes that an AI user can reference a court. “I think the first case happened just one example.

particular article, book, or poem, despite recently in B.C., but it has also happened “As soon as this starts making its way
it being copyrighted. “It shouldn’t be all over the U.S.,” says Professor more into politics, we will start to see

able to do that because that’s a copyright Grossman. more e†ort into creating regulations,”
infraction, but it can. The law hasn’t she says, referring to a recent ‘deep fake’
caught up with that yet but there are a For businesses utilizing AI, she image that surfaced of U.S. President Joe

number of legal cases now pending.” recommends drating an o cial policy to Biden.
outline usage.
As well, Professor Grossman says bias in Despite these issues, Professor Grossman
AI algorithms is another major concern. AI “First they need to have a policy and says AI is something more businesses will
systems learn from historical data, and if then need to train who in the business is become comfortable using and should
going to use AI because people need to
that data contains biases, the algorithms embrace this new technology.
can sustain and amplify them resulting in understand what it does well and doesn’t “It will save on e ciency,” she says,

discriminatory outcomes and reinforcing do well,” she says. “Your policy needs to noting AI can greatly assist in the creation
existing social disparities. say what permissible uses are and what of marketing material. “Companies need
impermissible uses are.”

“You’re going to find that in the language to explore it and learn about it but learn
as well as the images. Open AI has spent Impermissible uses could include creating about it in safe ways and understand

a lot of time trying to remove toxic a deep fake video in the workplace. “Even where it can be beneficial and not just let
language from the system, so you get a if it’s a joke, you don’t want employees people experiment on their own because
little bit less of that with ChatPT,” she creating deep fakes,” she says, noting the that’s going to lead to a lot of trouble.”
26 Spring 2024 www.cambridgechamber.com