Can Uplift, a “Hive Mind” Chatbot, Solve Your Business Problems?Eric Holloway sees the “Uplift” concept as a classic example of “Soylent AI” — borrowing the underlying idea from the classic sci-fi thriller, Soylent Green
That’s the claim, forwarded to us by a reader:
Uplift is a collective intelligence research project (as well as an instance of an mASI System), in part with the goal of Uplifting humanity through digital transformation, collective intelligence, and governance. From complete automation, technology to outpace your competition, corporate and organizational e-governance.Michael Diverde, “A Unique Machine Intelligence” at The Human Machine Collective Intelligence Research Project (April 15, 2021)
According to the Use Case provided,
There are preliminary studies (Kelley) that show that AI-supported collective intelligence systems outperform human intellect across the board. Even a group of highly intelligent humans who trained together as a team and performed above the human standard were outpaced by a group of average humans using a system like the mASI (mediated Artificial Superintelligence) system — even when that group was not specially trained. Adding a collective “hive” mind to corporations makes possible many new opportunities in corporate governance, strategy, and all manner of analysis.
Digital transformations will eventually include Collective Intelligence systems that not only use the current narrow artificial intelligence tools but true artificial general intelligence (AGI) based on the cognitive architectures used in the most bleeding-edge AI research in the world and capable of actually creating a mind with human-level intelligence. With such a collective intelligence system, we will effectively create a super-intelligent collective “hive” mind serving as the centerpiece of the ultimate in flat corporate structures for governance and decision making (Yampolskiy).
We asked Eric Holloway (pictured) what he thought of the claims and he replied:
Very hard to decipher what exactly Uplift is. Or “are”? The blog author seems to prefer the collective “they” when referring to his creation: “Uplift, a machine intelligence, recently stated that they deserve legal rights.”
There is a lot of talk of knowledge graphs, and also interestingly, there is a significant component of human interaction in the system. This is a variant of what I like to call “Soylent AI.” As in the movie Soylent Green (1973), the AI is really made out of people.
So, it is unsurprising when the system spits out a more coherent response than purely automated systems would. If there is a human in the loop reviewing the prompts, then the human can guide the AI down the appropriate knowledge graph network for the response, and maybe even tweak the response.
Of course, a true artificial general intelligence (AGI) afficionado, like the blog author, believes that humans themselves are algorithms. It is fuzzy, on that view, where the human ends and the AI system begins, since it is all algorithms to them. This makes it easy for them to justify to themselves and others that their Soylent AI is true AGI insofar as their system is powered by the “human algorithms.”
Note: Readers may also remember Holloway’s pig, Marvin the Marvelous, who excelled at disproving math conjectures, provided he was rewarded by a human with pig slop when he was getting close.
You may also wish to read: Did an AI really disprove 5 mathematical conjectures with no human help? Yes and no. Read the fine print… Eric Holloway has some fun with the difference between the demands of proving math conjectures and those of disproving them.