The tech industry's latest artificial intelligence constructs can be pretty convincing if you ask them what it feels like to be a sentient computer, or maybe just a dinosaur or squirrel. But they're not so good — and sometimes dangerously bad — at handling other seemingly straightforward tasks.

Take, for instance, GPT-3, a Microsoft-controlled system that can generate paragraphs of human-like text based on what it's learned from a vast database of digital books and online writings. It's considered one of the most advanced of a new generation of AI algorithms that can converse, generate readable text on demand and even produce novel images and video. [1]

Figure 1. Push to Open Up ‘Black Box’ of Language Models Amid Rapid Growth in AI Capabilities

Figure 1 shows they were trained on a large body of text and other media, these potent and resource-guzzling AI systems—technically known as “large language models”—are already being incorporated into customer service chatbots, Google searches, and “auto-complete” email features that complete your sentences for you. However, the majority of the computer corporations that created them have kept their inner workings a secret, making it challenging for outsiders to comprehend the flaws that could make them a source of false information, prejudice, and other negative effects. [2]

Competitive pressure to build the most eloquent or informative system — and profit from its applications — is one of the reasons that most tech companies keep a tight lid on them and don’t collaborate on community norms, said Percy Liang, an associate computer science professor at Stanford who directs its Center for Research on Foundation Models.

“For some companies this is their secret sauce,” Liang said. But they are often also worried that losing control could lead to irresponsible uses. As AI systems are increasingly able to write health advice websites, high school term papers or political screeds, misinformation can proliferate and it will get harder to know what’s coming from a human or a computer. [3]

One of many latest AI experimental fashions on the scene is Google’s LaMDA, which additionally incorporates speech and is so spectacular at responding to conversational questions that one Google engineer argued it was approaching consciousness — a declare that obtained him suspended from his job final month.

“I would not desire a textual content mannequin allotting medical recommendation or performing as a companion,” she stated. “It is good at that floor look of that means in case you are not studying carefully. It is like listening to a lecture as you are falling asleep.” [4]

References:

1. https://gadgets360.com/science/news/artificial-intelligence-gpt-3-microsoft-language-models-meta-ai-bloom-3167736
2. https://botgadgets.com/news/ai-researchers-push-to-open-up-black-box-of-language-models-amid-rapid-growth-in-ai-capabilities/
3. https://m.thelocalreport.in/ai-researchers-push-to-open-up-black-box-of-language-models-amid-rapid-growth-in-ai-capabilities/?amp=1
4. https://bestsaree.in/ai-researchers-push-to-open-up-black-box-of-language-models-amid-rapid-growth-in-ai-capabilities/

Cite this article:

Thanusri swetha J (2022), Push to Open Up ‘Black Box’ of Language Models Amid Rapid Growth in AI Capabilities, Anatechmaz, pp.252