AI Dilemma — Who’s Truly Responsible?
Beyond Algorithms: Unmasking the Human Factor in the AI Blame Game
AI hallucinations have been around since the 1950s, but it wasn’t until much later that people started talking about them. The term “AI hallucination” was first mentioned in research papers in 2000 during the Fourth IEEE International Conference on Automatic Face and Gesture Recognition. In a report published in 2022, the authors discussed the early adoption of this term from the field of computer vision, tracing its origin back to the 2000 publication.
In 2018, researchers from Google DeepMind brought the term “AI hallucinations” into the spotlight. As more AI tools emerged, the term gained popularity. Its usage reached its peak in 2022 when ChatGPT became publicly available, showcasing numerous examples of hallucinations to the general public.
AI can go wild
Let me tell you about some wild hallucination incidents that happened before! They weren’t exactly the talk of the town, but they sure left a mark. Picture this: it’s 2017, and Microsoft unveils its AI chatbot named Tay. Now, they thought it would be a fun and interactive experience, but were they in for a surprise!
Within just a day of its launch, Tay started spewing out some seriously random and downright offensive tweets. I mean, we’re talking about tweets that would make your jaw drop! Microsoft had no choice but to shut down Tay faster than you can say “oops.”
Lesson learned: even AI can have a bad day and go on a Tweeting rampage, or maybe we trained it wrongly?
News or Not?
I am here in Hamburg, watching the news in a language I don’t understand: but I do understand they are intervening an “expert” showing photo shoots for Press News about the War in Gaza claiming that are deep fake.
On all the media platform many questions arise about who we can trust: and honestly this is not a new topic! The Information system has always been considered as an Alternative Power entity, not always following clear ethic rules. It is not a problem related to Social Media, it is a clear problem related to Social Mindset and Culture.
In the past years our fears and suspects on the impartiality of Global Information Network was growing: now everyone is trying to shift the blame to the Social Networks, but is that really a thing?
Who is really shaping the Narrative? Facebook? Twitter/X? Or the News networks in our Countries with all the underlined Political implications?
I have to say that, for how deep is your education and critical thinking , it is not an easy feat to set out head straight and evaluate the situation.
And the list goes on: it looks like now, more than ever before, considering and confirm authenticity of the sources is a must.
Is AI to be blamed for all this mess?
When it comes to AI language models, it is essential to recognize that they are neither inherently good nor bad. The true issue lies in the improper education and understanding of the individuals who utilize these models for nefarious purposes or misguided objectives.
What kind of culture do we have to create and promote?
What regulations must be put in place at all levels (philosophical, ethical, technical, educational and political)?
Do we really need to have laws?
A computer can never be held responsible
Forty-five years ago, someone at IBM was already concerned about what we would be able to create and wrote those lines in the training manual emphasizing human accountability. Now, in 2024, in the context of AI, we have to consider that the ethical concerns are even greater when AI impacts access to public benefits.
So far, we have seen some little actions coming from governments around the world: the American AI Act and the EU Artificial Intelligence Act.
What is the real status today? What are the expectations?
If you want to learn more you can read for free this article from Medium.
This is only the start!
Hope you will find all of this useful. Feel free to contact me on Medium.
I am using Substack only for the newsletter. Here every week I am giving free links to my paid articles on Medium.
Follow me and Read my latest articles https://medium.com/@fabio.matricardi