Unanimous A.I., a startup based in San Francisco is using swarm intelligence in combination with AI algorithms to build AI that makes great decisions.
Swarm intelligence refers to the collective behavior humans often display when working in a group. CEO of The Unanimous AI, David Baltaxe told crowds at the Credit Suisse Asian Investment Conference that their technology is working.
“We have had great success in predicting and forecasting things like the results of sporting events, political events, and even the Oscars.” He went on to say their method of blending human intelligence with AI is performing better than traditional systems of artificial intelligence.
Swarm and AI accurately predicted Oscar wins
The tech from the company managed to accurately predict the Oscar winners in all but one of the total 16 major categories. Unanimous AI describes their work saying “Our technology is called Swarm AI®. It’s a set of algorithms and interfaces modeled after swarms in nature, empowering populations of online users to amplify their collective intelligence by thinking together in closed-loop systems.”
The company offers private research services to a range of industries. They promise they can provide accurate and insightful intelligence on across topics including market research, product development, and sales forecasting.
Tech available for hire via subscription
The company sells this research through a subscription service called Swarm Insight®, it promises to build and deploy swarms of real people to gain insights into any topic, it then combines this human intelligence with its cutting-edge AI algorithms to provide predictions and insights.
Swarm AI uses real human decision making which requires a large community. People are invited to be a part of the company's Swarm Sessions.
At these sessions, information is proposed to the group who make decisions and predictions which are then used in combination with AI to generate the final outcome. “When groups of people get together, the output is much more accurate and satisfactory to all participants because everyone is participating,” Baltaxe states.
Unanimous A.I. are motivated to keep the humans at the forefront of the discussion and development around AI. “AI systems may be able to have great control over our lives, but [if the systems exist externally] we may not be able to comprehend why they come to certain decisions,” says Baltaxe.
Humans can badly influence algorithms
The question of how humans influence AI algorithms is an interesting and ongoing debate. All AI is developed by humans and therefore has their mark on it, no matter how independent the algorithm is designed to be.
While Unanimous A.I. is working within the relatively low-impact field of predictions, humans can have a serious impact on AI development.
Humans innately have prejudice and these can be inadvertently built into AI. Many AI models rely on huge amounts of data that allow it learn and process new information.
"Tay" went from "humans are super cool" to full nazi in <24 hrs and I'm not at all concerned about the future of AI pic.twitter.com/xuGi1u9S1A— gerry (@geraldmellor) March 24, 2016
It was previously assumed the bigger the initial data set, the more fair and broad the results will be. However recent examples have shown that this is not always the case with AI quickly learning our bias and prejudices.
As the AI field rapidly increases scientist, engineers and thinkers all have a responsibility to continue to interrogate our ethics and intentions.