John Oliver ranted about AI on 'Last Week Tonight'
During a recent episode of his popular Sunday night talk show on HBO, "Last Week Tonight," Oliver criticized the rapid adoption of artificial intelligence and machine learning models, specifically the Microsoft-financed chatbot ChatGPT.
"The problem with AI right now,” Oliver said, “isn’t that it’s smart. It’s that it’s stupid in ways that we can’t always predict. Which is a real problem, because we’re increasingly using AI in all sorts of consequential ways.”
Such comments are unfair. AI is useful in many different sectors, including education and finance, and large language models such as ChatGPT and Google's LaMDA and Bard are impressive in the number of tasks they can accomplish.
For example, the ability to quickly generate relevant information is helpful when you’re building content like the “help” function on a website.
Opinions in your inbox: Get exclusive access to our columnists and the best of our columns.
AI also is useful in assisting with customer service. People want to feel like they’re interacting with a human while being taken care of quickly. AI helps achieve that experience, and in a world where so many of our interactions are conducted remotely, this matters.It’s also important to understand that the models are evolving. Like Siri and Alexa, they’ll get smarter as they receive more content from wider audiences.To be fair, Oliver made some valid points. The proliferation of AI models does highlight challenges that more people should be talking about. For example, ChatGPT and other AI programs are great at cleaning code. But with developers from all over the world using ChatGPT to clean up code, they're actually giving away intellectual property by feeding the chatbot.
AI can be misused, but there are solutions
Another huge problem is black box AI technology, which lacks transparency in how it makes decisions and in turn lacks accountability for churning out misinformation. Now, even someone with limited technical experience can download widely available black box AI models from the internet, like XGBoost, and figure out how to train them.
But they’ll have no clue why the model made the decisions it makes and whether bias and false information tainted the results.
In his 28-minute rant, Oliver warned against the dangers of black box AI. He noted: “AI systems need to be explainable, meaning that we should be able to understand exactly how and why an AI came up with its answers.”
He failed to mention that there is a more transparent and safer alternative to black box systems. Silent Eight has established itself as a market leader in AI by developing pioneering white box solutions in identifying financial crime for leading banks such as HSBC, Standard Chartered and First Abu Dhabi.
Is Elon Musk a risk to US security? Pentagon's reliance on him grows despite his behavior.
Our flagship platform, Iris, is fully explainable and tells you how and why it made a particular decision. Similar models exist in a host of industries beyond finance.
You can find complex, pioneering models with clear explanations for why a certain result is produced. But they aren’t widely available or easy to download from the internet. These tools are methodically crafted, monitored, tuned, measured and tested by skilled mathematicians, data scientists, subject matter experts, and risk and compliance analysts.
Smart machine learning and AI technology are valuable technologies that are integral to the modern world. While you should never use black box AI to make decisions that have legal consequences or serious ramifications, that doesn't mean you can’t apply complex, cutting-edge and transparent solutions to improve the decision-making process in critical industries.
I don’t think the general public has anything to worry about in terms of black box solutions. ChatGPT is an impressive technological advancement that will be helpful to a variety of industries in incalculable ways.
Tell that to John Oliver.
https://bit.ly/3K0LOUE
https://bit.ly/3FGa1x3
0 Comments