If you wish Siri were a little more natural or communicating with technology were smoother, these five companies are helping fix the problem across gaming, law and learning
For most, Artificial intelligence (AI) – loosely defined as the ability of machines to mimic human intelligence – feels like a known entity borne out of sci-fi and pop culture. However the reality of AI the burgeoning startup sector growing around it is perhaps far more exciting than anything ever dreamed up in Hollywood.
AI is a diverse field, mostly dominated by efforts in machine learning that use algorithm frameworks to gather unprecedented amounts of actionable information. Within that field is the subfield of deep learning, which might borrow from fields like computer vision and image recognition to analyze data (text, images, video, speech, music, etc.) in order to properly identify something. Think about a computer seeing millions of pictures of a dog. It begins to understand that the right combination of tails, snouts, legs, etc make up a dog. Eventually it can be shown only a part of picture of a dog and still correctly identify it as a canine even though it is missing some information. Like a human who makes an educated guess based on past experience, the AI can reasonably assume based on seeing so many pics of dogs, that the image it is seeing has enough potential to be properly categorized as a dog. That is where the majority of artificial intelligence work is right now. It is complex, but it is still cold and mechanical.
Yet, figuring out the logic behind an algorithm’s decisions is not obvious. But a number of startups are trying to make AI a little more warm and less mysterious.
As previously fictional concepts like self-driving cars enter our lives, great minds and aspiring entrepreneurs are trying to smooth out interaction between people and machines, be they apps or coaching programs.
In some ways, the integration of AI is giving people time to focus on things other than stats and material like legalese, so the humans using it also have the potential to become a little more, well, human. Here are a handful of those companies from Silicon Roundabout in London who want to make AI feel a little less robotic.
1. Weave.ai (London)
Founded in London in 2015, Weave.ai wants Siri to feel a little more human. The company wants to get AI to better explain its decisions and evaluations to the humans that use it, adding a bit of transparency to artificial intelligence that Co-founder and CEO Rodolfo Rosini calls “explainable artificial intelligence.” They have already raised $200,000 thanks to seed investors like Techstars.
“Today there’s no way to trust the algorithm,” Rosini told the Financial Times. “If a computer tells a doctor you have cancer, the doctor doesn’t know why. You can’t ask it why it thinks that, it could be a false positive.”
“Our system would work for anything where the user has to communicate with an AI.”
2. PROWLER.io (Cambridge)
Breaking out with its first seed funding at the end of September, Cambridge-based gaming startup PROWLER.io is developing new algorithms that would enable “more natural in-game behavior” in their words after substantially improving the quality and speed of AI decision-making.
“With so many games now relying on extended playing time to generate revenue, it’s more important than ever for developers that games don’t become boring or repetitive,” PROWLER.io CEO Vishal Chatrath has said. “Sophisticated AI and machine learning is the best way to create and deliver that experience, and can reduce cost at the same time. It’s a win-win.”
Backed by Passion Capital, Infocomm Investments and Amadeus Capital, the company already has £1.5 million ($1.94 million) in the bank. The company was founded by Chatrath, CTO Dr. Dongho Kim and Aleksi Tukianen.
3. Bloomsbury AI (London)
Incubated by Entrepreneur First (EF) in London, Bloomsbury AI presented at EF’s 2016 pitch for its sixth cohort on 6 September 2016. Their goal is to use AI to smooth out the learning curve for aspiring programmer, teaching them to learn and model more quickly.
Their slogan is “Enabling computers to understand human language so humans don’t have to understand computer language.”
“With colleagues from UCL and UW, we created the first company providing a Natural Language Programming system,” in Co-founder Guillaume Bouchard‘s words on his LinkedIn profile. “The company provides to its customers a ‘machine reading’ web service that reads textual data and can answer automatically some questions about this text, as if you were doing a school exam.”
4. Luminance (Cambridge)
It’s hard to find someone who accuses lawyers of being human (sorry), but one spinout from the University of Cambridge announced a $3 million investment on 27 September 2016 to speed up the analysis of legal documents, speeding up the process by learning how lawyers think along the way (which will unlock mysteries for a lot of people actually).
As Luminance Founder and CEO Emily Foges has said, “AI won’t replace lawyers, but can it make law fun again?”
So in some sense, despite the title of this article, this startup might actually make select humans seem a little less robotic. Rather than looking into the minutiae of contracts, Luminance will pick up on anomalies automatically. Therefore, lawyers will be a little freer to focus on actual negotiations and on their clients.
“This will transform document analysis and enhance the entire transaction process for law firms and their clients,” Foges has said. “Highly-trained lawyers who would otherwise be scanning through thousands of pages of repetitive documents can spend more of their time analysing the findings and negotiating the terms of the deal.”
5. Eblur (London)
Also incubated by Entrepreneur First in London, the gaming-focused Eblur also pitched at EF on 6 September 2016. Calling their solution Feedless, it acts an AI coach for eSports and gaming. It talks to you while you play and suggests counters mid-game. It even comes with seven different voices.
They say in their FAQ that their gamers won’t get VAC banned for using their tech, and that Feedless “works the same way as a screen recording app.”
As one beta tester put it in his video review below, “Simplified, it’s like having an adolescent Siri with a different brain at your beck and call.”
The company was co-founded by Max Colvin and Joao Seabra.