Generative AI: 7 Reasons Why You Can't Afford to Stay Behind, But You Also Need to Be Careful at This Point.
Tuesday, March 26, 2024
AI is like your child : You love it, but you cannot trust it yet..
If you believe AI is just hype and you can live and work without it for the next 10 years, you may be replaced by someone who knows how to use it. Would you have a job if you didn’t know how to use the internet, a computer, or office suite? Probably not.
Let us guide you on why AI is a must-have first, and second, how great it already is, not just in theory, but in practical life.
Generative AI has found real-life applications from day one, and it's essential to see how this powerful technology can truly make our lives easier and better.
First, let's get excited because you may belong to the group of laggards who believe this is just business as usual. You need to know one thing, and you need to know it right now: people who believed the internet was just hype are all out of business, and people who thought the iPhone was just a gadget do not exist anymore.
Why you can’t overlook what is happening.
The impact of the internet on businesses has been profound, with many traditional models being disrupted and even eradicated. Here are some industries and businesses that have felt the effects:
Retail
Brick-and-mortar retail stores faced challenges as online shopping gained popularity, leading to struggles in competing with e-commerce giants like Amazon.
Print Media
Newspapers, magazines, and other print media have seen declining readership and ad revenue as online news sources and digital advertising platforms have become more prevalent.
Travel Agencies
The rise of online travel agencies and booking platforms has allowed consumers to book flights, hotels, and vacation packages directly, reducing the need for traditional travel agencies.
Additionally, industries such as video rental stores, postal services, music stores, traditional advertising agencies, and bookstores have also experienced significant disruption.
Therefore, if you wish to remain viable in your field over the next decade, it's crucial to educate yourself on what AI can and cannot do, ensuring your continued relevance.
Why you should be super excited about generative AI
Let's delve into the incredible discovery of generative AI and why it holds such significance. As a user, you're thrilled to witness its ability to comprehend questions and provide detailed answers, which is truly mind-blowing. You find yourself utilizing it once or twice daily for knowledge queries.
Generative AI boasts real-life applications that you can begin utilizing today, many of which may be unfamiliar or challenging to access. Prepare to be amazed by its capabilities.
Gerenative AI Text Summarization
Generative AI can condense lengthy emails or email conversations into succinct bullet points. Essentially, this means AI can be trained on your data to perform on a small scale what it accomplishes on a larger scale using general knowledge from the internet.
Consider this—it's truly a game-changer:
• Just returned from holiday with 500 emails? PLEEEEAASSE summarize what I missed for me.
• Oh no, another conversation spiraled out of control in my mailbox? Summarize it for me."
Generative AI Meeting Summary
With 37 participants and a 2-hour-long PowerPoint presentation, you decided to join but found it difficult to follow what was happening. Don't worry; we've got you covered here as well
Generative AI can take recordings of meetings and summarize them, including action items and decisions. While this may sound like science fiction to some, it's already a reality. But how does it really work? Do you need to check it out, or can you trust it as is? These are the points we'll discuss.
Generative AI Language Translation
Alright, let me tell you. You've been using Google Translate for 10 years, and your content with it. But the truth is, it has been and still is the best option available until now.
Now, imagine being able to translate text 10 times better than Google Translate. With a tool that comprehends context and enhances its performance over time—yes, it improves over time with your feedback
Google Translate is one of the ancestors of generative AI. It achieved this by reading the same book in multiple languages and then deciphering how to translate existing sentences based on manually translated content.
Gerenative AI Video Creation
You can create a video in any language and have it automatically dubbed in Hungarian, Chinese, and more. This opens possibilities for educators like us, or managers, to send videos in any language. Believe it or not, not everyone speaks English fluently or uses it as their mother tongue (I hope you knew that, otherwise sorry for the news). To prove this point, I advise you to go on holiday to Japan, China, Italy, or France...
Writing Support
You can write with up to 200 mistakes, including 50 grammar errors, lack clarity in your text and ideas, and still have AI rewrite it for you. This applies to emails, reports, blogs, and even books. You only need to provide the basic text, and AI will not only correct but also enhance it. This is excellent news for non-native speakers and those lacking confidence in their spelling and grammar.
How did it learn all this stuff?
The internet stands as the largest source of human knowledge today. While it's vast and rapidly expanding, it's important to note that it's not comprehensive, as many books have yet to be digitized. However, despite its size, it's not necessarily the best source of knowledge.
Anyone can download the internet.. really
Yes, you read that right. Some businesses or non-profits crawl the internet, jumping from link to link until they've indexed all pages and save the internet's text in a vast storage. This data isn't too bulky, as it's primarily in text form and accessible from home as well.
Internet Knowledge can be misleading
Michael Jackson is not dead, and he is a doctor in Kentucky now.
The internet is rife with misinformation, outdated information, biased information, and toxic content. For example, if you search across all pages about Michael Jackson, you'll find some claiming he is still alive while others state otherwise.
Another challenge with the internet as a source of information is the ambiguity of the data. At Knil, we studied this ambiguity and discovered that there can be over 500 different attributes required to make a named entity like 'Michael Jackson' unique
Internet is not reliable
Internet facts are not consistent, even on the same topic. Over the past years, we've come to recognize the concept of fake news, but even simple static content can be incorrect or outdated.
When you read 500 pages on the same exact topic, you'll encounter many similar patterns, which is beneficial for learning since we also learn through patterns. However, what do you do with the differences, even on facts? Who can you trust—the most popular source, the most recurrent fact?
You can learn to kill, hate and more
The internet is replete with toxic, racist, and dangerous content. You can learn how to become a terrorist, build a bomb, bury someone alive, and more—all this information is readily available, formatted in words, images, and videos like any other content. So, how can a robot make any difference in determining whether content is harmful or not?
So what is scary or dangerous about it
Tobacco companies are sued every year by smokers with lung cancer, video game makers by addicted players, and now Facebook by addicted users. But who is responsible—the cigarette or the smoker, the game or the player?
Great power comes with great responsibility, and we find ourselves right in the middle of it.
Confidence in wrong facts with AI
As we learned before, AI learns from non-exact sources, so it can make mistakes. None of these mistakes are intentional, but they can be particularly dangerous because they are confidently wrong and often involve logical errors. AI produces mistakes that you can trust, and with style
- It is currently quite bad at math.
- It is making coding errors.
- It is using outdated information from 2021.
- It is making mistakes in reasoning.
Plagiarism at scale with AI
You can commit plagiarism with Taylor Swift songs and J.K. Rowling books in the morning and start distributing them in the afternoon.
Cheating at an exam with AI
The advent of new technology has posed a new challenge for children educators, as children have come to understand the power of technology and utilize AI to write their essays for them.
This means they can pass exams without attending or learning anything.
What next..
So, what can we do about it? Tune in to the next episodes to learn how prompting, privacy and guardrailing can help us overcome these challenges.