up:: π₯ Sources
type:: #π₯/π°
status:: #π₯/π©
tags:: #on/articles
topics:: π€ Artificial Intelligence, ASI, AGI
Author:: spiceworks.com
Title:: What Is Super Artificial Intelligence (AI)? Definition Threats and Trends
URL:: https://www.spiceworks.com/tech/artificial-intelligence/articles/super-artificial-intelligence/amp/
Reviewed Date:: 2023-01-09
Finished Year:: 2023
What Is Super Artificial Intelligence (AI)? Definition, Threats, and Trends
Highlights
Superintelligent AI may lead to the singularity
- ASI will be one of the finest and possibly the last inventions that humans would ever need to make as it will constantly evolve to become more intelligent.
- Superintelligence will accelerate technological progress across various fields such as AI programming, space research, discovery and development of medicines, academics, and plenty others.
- ASI may further mature and develop advanced forms of superintelligence that may even enable artificial minds to be copied.
- ASI, in the future, may also lead to technological singularity.
- Loss of control and understanding
- The weaponization of super AI
- Failure to align human and AI goals
- Malevolent superintelligence
The successful and safe development of AGI and superintelligence can be ensured by teaching it the aspects of human morality.- The danger of nuclear attacks
- Ethical implications
The problem is humanity has never agreed upon a standard moral code and has lacked an all-encompassing ethical theory.
Super-intelligent AI can have serious ethical complications, especially if AI exceeds the human intellect but is not programmed with the moral and ethical values that coincide with human society.
Advancements in multimodal AI
Deep learning algorithms have trained the underlying models on a single data source. For example, an NLP model is trained on the text data source, while a computer vision model is trained on an image dataset. Similarly, an acoustic model uses wake word detection and noise cancellation parameters to handle speech. The type of ML employed here is a single modal AI as the model outcome is mapped to one source of data typeβtext, images, and speech.
On the other hand, Multimodal AI combines visual and speech modalities to create scenarios that match human perception. DALL-E from OpenAI is a recent example of multimodal AI-generated images from texts. Googleβs multimodal AI β multitask unified model (MUM) β helps enhance the user search experience since the search results are shortlisted by considering contextual information mined from 75 different languages. Another example is NVIDIAβs GauGAN2 model, which generates photo-realistic images from text inputs. It uses text-to-image generation to develop the finest photorealistic art.
AI-generated inventions
AI is already powering inventions in almost every domain, from creating music, art, and literature to developing scientific theories. Recently, DABUS, an artificial inventive machine, came up with ideas for two patentable inventions. While the first invention relates to a device that attracts attention and is helpful in search and rescue operations, the second is a type of beverage container.
Moreover, it has been observed that the DABUS machine also has an emotional appreciation for whatever ideas it conceives
up:: π₯ Sources