AI Assistants are rapidly learning to provide any sort of answer in moments, often without citing verified or real sources. Even simplistic queries can deliver sophisticated answers that require little or no creative thought, bypassing the process of learning and understanding the subject. If instant gratification replaces analytical thinking and human skills, we run the risk of becoming shallow thinkers, following predictions made in 2010 by Nicolas Carr, author of 'The Shallows': This Is Your Brain Online.
Surfing the Internet shortened our attention spans. But we still had to think enough to search, ask questions, to critique the sources, and assemble results. AI Assistants take the next step, nearly eliminating the need to concentrate and learn, delivering instant gratification, pushing us further into the shallows. Note 1 provides a very short list of creative and research tasks that are effectively bypassed by the earliest of the new generation of AI Assistants. There is no time to waste in recognizing and addressing the impacts of AI Assistants.
Internet Commerce Impacts
Why would web users make the effort to search for information when they can get an answer in moments? Government, retail, news, medical, academic, and many other types of sites will be less frequently visited for the wider views and choices of any given topic and therefore will lose value and revenue.
Content Creator Impacts
Content creators in every field are directly impacted, including writers, artists, musicians, lawyers, teachers, counselors and more. AI Assistants can compose virtually anything that a user requests. The risk to devalue and pirate genuine creative and deeply learned human talents is palpable. Rights challenges are inevitable.
Legal Impacts
Copyright laws have no historical basis for the scale of derivation and mimicry that can take place when AI Assistants call upon massive historical data to compose answers. Lawsuits over AI provenance, piracy, imitation, and invalid advice are just getting started (see Note 2). AI Assistant companies will be obliged to prove the sources from which their tools drew responses. Lawmakers will be inundated with demands for clarity.
Cybersecurity Impacts
IT Professionals may not fully appreciate the speed at which this change is taking place, and the impact these changes will have on cybersecurity. ChatGPT for example, will write code, and there are concerns that it is being used to create malware. Extensive sampling of public data can allow AI Assistants to create interactive audio and video conversations that can effectively mimic any person will eventually cause serious damage to validation processes that rely on traditional authentication steps including voice and visual recognition. Visual “Deep Fakes” that are increasingly difficult to test by legacy methods have been used to manipulate news stories and to foment conspiracies and dissension. AI voice impersonations have already successfully contributed to costly cyber security attacks (Note 3).
Make a “Real” Intelligent Action Plan for Your Organization NOW
In a matter of weeks, ChatGPT has sent massive ripples through the Internet community and forced directional changes among information providers such as Microsoft and Google (Note 4). AI Assistants cannot be dismissed, therefore they need to be managed by understanding their impacts. It’s time to think of ways for them to be useful as adjunct information sources. More ripples are coming. My recommendations:
- Bone up on stories of dystopian futures where AI caused benefits or disasters. If you are not a Sci-Fi fan, it’s time to admit that those futuristic stories have been coming true for decades. Take time to read about the technical and psychological impacts that past authors have imagined. Refer to Note 5 for some recommended reading.
- Require three credible information sources, checked for authenticity, for any new business process proposal and any new recommendation. Require concept champions and their managers to demonstrate knowledge and understanding, not just unqualified bright ideas. Require verifiable examples. AI Assistants are prone to “hallucinations” in which they invent “facts and references”! Are you surprised? See Note 6.
- Never assume that intellectual property rights are secured for any collateral: written, audio, visual, etc. Review copyright law with a human counsel, and require proof of provenance and licenses. Imitation may be flattery, but AI Assistants may borrow too much (Note 2).
- Review all of your organization’s security and authentication procedures that rely on behavioral criteria including location, device, network, face, voice and fingerprint. Never assume that AI Assistants cannot replicate all of these factors. The entire concept of multifactor authentication must be reviewed. NOW!
AI Assistants will not disappear, and they will get better at doing work that leads us to shallow thinking. Since they have a means to learn, take the opportunity to use them as collaborative tools with vast information and the naivete of a child. Don’t rely on them to replace your ability to think: it’s a self-fulfilling prophecy if you do.
------------------------------
Note 1. Tasks that emerging AI Assistants will deliver in a matter of seconds to a few minutes.
- How would you write a term paper in school? Go to the library, talk to scholars, immerse yourself in the topic, history and effects? Or simply ask an AI Assistant to write the term paper for you?
- You want to express love to your fiancée with a poem. Do you think about your relationship with empathy, pick a style, read the works of poets, then say what you feel, or do you just ask an AI Assistant to write it for you?
- You need to critique a famous painting. Do you spend time looking at it from different points of view, drawing on your memories and motivations and the comments of others? Or do you just ask an AI Assistant to write an essay?
- You need an image. It may be a graphic, or some fine art. You can search online collections navigate the licenses, or make and tweak your own, or commission an artist. Or you can ask a graphical AI Assistant to generate portfolios from your goal descriptions and your own image samples.
- You need legal advice. You can call a lawyer and pay for a consultation, search/pay for templates, with explanations as to how, when and why to use them. Or you can ask an AI to to write a legal document based on your circumstances.
- You need to translate a book. You can use a dictionary, and enter words and phrases piecemeal into translator apps. Then check that meanings are on track. Or you can load the book into an AI Assistant and it will do the job for you, even correction transcription errors. You can also ask it for alternate
interpretations. - You need a music theme for your brand or vlog. It might be a solo instrument, band or even a full orchestra. Will you search through sound and music libraries, maybe hire a composer? Or, simply describe your desires to an AI that can generate anything you ask.
- You want to tell your child a personalized bedtime story. Do you make up stories on your own? Do
you pay for custom books and songs? Or will you just give some details to an AI Assistant and ask it to compose a story in the style of a favorite author? - The family is visiting and you want a movie that everyone can enjoy. Ages range from 5 to 60; every generation is in the house. Will you discuss ideas? Go online to look at reviews, lists posted by other people? Give some thought before making a choice? Or will you just ask an AI Assistant to make a recommendation?
- A movie studio wants to use a popular actor in a new role, but the actor has aged, is injured, or is deceased. Will they give the role to a new actor with a different face and voice? Or will they use a new class of AI Assistants that automatically map the original actor’s face and reconstituted voice, with little or no need for editing?
Note 2. Legal Impacts - Copyright challenges for AI Assistant generated content
Who Ultimately Owns Content Generated by ChatGPT and other AI Platforms?
Exploring the Legal Minefield of ChatGPT and Intellectual Property Rights
Art-istic or Art-ificial? Ownership and copyright concerns in AI-generated artwork
Who Owns DALL-E Images?
AI Music Copyrights
Even if a tool like ChatGPT can reject an overt request to violate copyrighted
materials, it can still be directed to mimic those materials.
OpenAI’s ChatGPT May Be Rejecting Prompts That Violate Copyright
Note 3. AI Impacts on Cybersecurity
Note 4. Chaos caused by the introduction of AI Assistants
Google vs. ChatGPT: Here’s what happened when I swapped services for a day
Will Conversational AI Kill Web Traffic?
Microsoft is making a multibillion-dollar investment in ChatGPT
Alphabet shares dive after Google AI chatbot Bard flubs answer in ad
Lecturers urged to review assessments in UK amid concerns over new AI tool
ChatGPT Wrote My AP English Essay—and I Passed
ChatGPT bot passes law school exam
Disney’s neural network can make actors younger or older
Note 5. RecommendedSci Fi
Humans have forgotten mathematics, including even the rudimentary skill of counting.
The Feeling of Power
Having made machines that perform all menial or routine labor, people lose touch with their human experience and regress both socially and intellectually as a species.
Twilight
An unexpected first contact experience that sort of follows the prior author’s premise.
They’re made out of Meat!
A.L.I.E's core command was to make life better for humans.
The100
Note 6. AI Assistant hallucinations, misbehaviors and lies
The creativity programmed into large language models allows AI tools to invent fictitious results, and wrong assumptions.
Hallucinations, Plagiarism, and ChatGPT
John Girard, Lionfish Advisor Team
CyberImbiber Tech Advisor LLC
Cover image by Starline on Freepik
Brain image by vecteezy