.
Subscribe to our Voices Dispatches email and receive a comprehensive summary of the top weekly opinions.
Please register for our complimentary weekly Voices newsletter.
Leading computer experts predict that in the future, various artificial intelligence units will have the ability to collaborate and exchange data, similar to the Borg in Star Trek.
The Borg are made up of cybernetic beings in a science fiction television series that function by connecting to a collective hive-mind called “The Collective.”
Researchers from Loughborough, Yale, and the Massachusetts Institute of Technology predict the rise of “Collective AI”, where multiple units with the ability to learn and acquire new skills, come together in a network to exchange information.
The group revealed their purpose in the publication Nature Machine Intelligence.
The researchers stated that while the antagonists in Star Trek or the Replicators in Stargate SG-1 are highly advanced machines, the effects of Collective AI are expected to be more beneficial.
Dr Andrea Soltoggio, lead researcher at Loughborough University, stated that in this newly formed group of AI systems, any new information learned by one unit can be shared with all other units. This bears a striking similarity to the advanced abilities displayed by sci-fi characters such as the Borg from Star Trek or the Replicators from Stargate SG-1.
Over the course of thousands of years, human understanding has gradually expanded due to communication and collaboration. It is predicted that similar patterns will arise in future societies of artificial intelligence, where democratic and cooperative collectives will be established.
This skill enables a group to be highly adaptable and reactive when facing new surroundings, obstacles, or dangers. This is because every piece of new information can be exchanged and incorporated into the group’s overall knowledge.
”
For instance, in the context of cybersecurity, if one artificial intelligence unit detects a danger, it has the ability to rapidly exchange information and initiate a coordinated reaction – similar to how the human immune system defends the body against external intruders.
”
This may also result in the creation of emergency response robots that can rapidly adjust to the situation they are sent to, or individualized medical assistants that enhance health results by combining advanced medical expertise with specific patient data.
“The possibilities for use are extensive and thrilling.”
Nevertheless, the specialists recognize that Collective AI may also entail some dangers, such as the quick dissemination of information that could be considered unethical or illegal.
However, they also mentioned that AI units can remain secure by consistently pursuing their own goals and maintaining their autonomy from the group. This leads to what Dr Soltoggio calls “a democracy of AI agents,” which greatly decreases the chances of a few large systems dominating AI.
According to the researchers, the AI Collective sets itself apart from existing large AI models like ChatGPT by possessing superior abilities for lifelong learning and knowledge sharing.
According to their statement, ChatGPT and other similar models acquire their knowledge mostly through intense training sessions that require a lot of energy. After the training is completed, these models are not able to continue learning.
According to Dr. Soltoggio, the prevailing, costly, non-transferable, and short-lived AI models will likely be replaced by sustainable, progressive, and collaborative AI units in the future.
He proceeded by saying, “The growth of human knowledge has occurred gradually over thousands of years through the means of communication and collaboration.”
“We anticipate that similar patterns will emerge in future societies of artificial intelligence entities, which will function as democratic and cooperative collectives.”
The US Defence Advanced Research Project Agency (DARPA) provided funding for their research.
Source: independent.co.uk
