The State of AI 2020 Report launches featuring Signal AI Research

LONDON / NEW YORK  — 1st October 2020 — Launched today, the third annual State of AI Report, co-authored by Air Street Capital and RAAIS founder Nathan Benaich and UCL IIPP Visiting Professor Ian Hogarth in collaboration with more than two dozen key industry partners, identifies this year’s top trends in AI, and predicts what’s coming next. The State of AI Report is a compilation and analysis of the most interesting insights in AI, which aims to trigger an informed conversation about the state of AI and its implication for the future and is widely seen as the most comprehensive of its kind.

Signal AI contributed to the report sharing its research on how its proprietary applied AI fuels both Reputation Perception and ESG (Environment, Social and Corporate Governance) insights, in turn allowing clients to better understand risks and opportunities in their reputation and ESG performance.

 

 

Co-Founder and Chief Data Scientist Dr Miguel Martinez, said:

“The global State of AI Report is the leading annual tome in the industry, hugely respected by academics, practitioners, business and political leaders alike. We are thrilled to have contributed to this year’s report alongside some other pioneers in the applied AI space, from leading global universities to other cutting edge technology companies.”

The report, now in its third year, covers the latest developments in Research, Talent, Industry and Politics. It includes contributions by over 25 companies and has been reviewed by a range of experts in the AI sector, including Jack Clark, Jeff Ding, Chip Huyen, Rebecca Kagan, Andrej Karpathy, Moritz Müeller-Freitag, Torsten Reil, Charlotte Stix, and Nu (Claire) Wang.

 Some of the trends the report highlights are: 

1) Natural language processing (NLP) is clearly a pivotal aspect of AI now. The report shows clear traction of NLP in Research, Industry, Funding and Policy, however there is a correlation in cost of computing power, concentrating the most exciting natural language processing (NLP) research among a few big tech-funded labs led or funded by Google, Microsoft, and Facebook. This carries implications for innovation outside of these well-funded labs with successful smaller startups focused less on first-principles research and more on applications that identify opportunities that existing ML can solve for.
• The report identifies the outrageous computational, economic and environmental costs of even incremental improvements in model performance. For example, the Microsoft-backed San Francisco lab OpenAI released its GPT-3 transformer-based language model that used 175 billion parameters in training – a 117x increase over last year’s GPT-2. GPT-3 may have cost $10m to train.

2) 85% of AI research papers do not publish code: Open sourcing AI research is important for accountability, reproducibility and driving progress in AI but in reality, the vast majority of AI research papers are behind corporate lock and key.

• Transparency has been slow to take hold: In 2016 about 90% of research did not publish code alongside their papers. Academic groups are traditionally more likely to publish code than industry groups.
3) AI holds the keys to drug discovery. Biology experiences its “AI moment”: This year signaled a new era for AI-first drug discovery startups with mega-rounds and first-ever AI drug in clinical trials. In spite of pharma doubts, AI-powered life science companies also went from being a speculative endeavour to something rapidly industrialising and scaling:
• UK-based startup Exscientia became the inaugural AI-first pharma startup to produce a drug candidate to enter clinical trials, with a treatment for OCD, while massive funding rounds for biotech startups such as Insitro and Recursion highlighted the maturation arch of AI-first biology companies.
• US startup Viz.ai became the first deep learning-based medical imaging product to qualify for Medicare & Medicaid reimbursement, cementing its official recognition by the US healthcare system.
• UK startup ZOE is running the world’s largest clinical study into COVID-19, analyzing symptoms from over 4 million contributors with a view to discover new symptoms, predict infection hotspots, and eventually, to use AI to predict COVID-19 without a physical test.
4) Next arms race? Explosion of AI-powered military: With another wave of countries declaring national AI strategies, more governments doubled down on the military adoption of AI technology.
• The US Air Force’s AlphaDogfight demonstrated how cutting edge techniques used to win in a game environment inspired by war can rapidly migrate to a military context. The US Secretary of Defense targets 2024 for real-life AI vs human dogfight.
• In response to US DoD activity and investment in US-based military AI startups, it’s possible that we will see a wave of Chinese and European defense-focused AI startups raising over $100M over the next 12 months.
• The rapid advance of AI technology by both industry and government creates the need for protections to prevent triggering a potential arms race as major powers compete for AI-powered dominance. The trajectory of AI research to stay private could pose national security risks if proprietary technology winds up in the wrong hands.

Nathan Benaich, co-author of the report, said:“Despite fears of an AI winter, this year’s report reveals the extent of innovation and progress behind the scenes. Many of these AI developments are  powered by the massive computing infrastructure that is increasingly in the hands of big tech companies. We need to think carefully about what this means for the future of the field across both industry and academia."

 

At the same time, the report finds more and more AI-first startups implementing the core ideas that emerge from AI research, creating highly-valuable products for significant problems in industry. I’m particularly optimistic about the significant impact that AI is starting to have in the life sciences as biology and medicine become large data domains ideal for AI applications. We’re on the brink of being able to decode a lot more about our health and revolutionise how we treat disease.”

 

Ian Hogarth, co-author of the report, said: 

“This year has seen a step-change in the application of AI in a military context, with the US being the most active player. The US Department of Defense is increasing its spending on AI-related projects, DARPA has organised virtual dogfights where an AI defeated a human pilot, and a number of AI defense startups have raised very large funding rounds. This risks triggering a proportional investment from other countries and amplifying great power conflict. We hope that a major investment in AI Safety research and appropriate new international legislation around the governance of AI will help steer us away from a literal arms race."

 

This was also the year that concerns about facial recognition and AI based decision making algorithms went mainstream, with a series of wrongful arrests and the exam grading fiasco in the UK. The reaction was international with lawsuits against the operators of facial recognition systems in the US, China and Europe. Lawmakers are now scrambling to find appropriate ways to legislate the use of machine learning and more thoughtful approaches are starting to gather steam.”

 

About The State of AI Report:

The State of AI Report is a compilation and analysis of the most interesting insights in AI, which aims to trigger an informed conversation about the state of AI and its implication for the future. Now in its third year, the report features invited contributions from a range of well-known and up-and-coming companies and research groups. The authors declare a number of conflicts of interest as a result of being investors and/or advisors, personally or via funds, in a number of private and public companies whose work is cited in this report.

 

Ian is an angel investor in Chorus.ai, ComplyAdvantage, Disperse, Faculty, LabGenius, and PostEra. Nathan and Air Street Capital are shareholders of Graphcore, LabGenius, Niantic, ONI, PolyAI, Secondmind, Tractable, and ZOE.

 

About Nathan Benaich:

Nathan is the Founder and General Partner of Air Street Capital, a venture capital firm investing in AI-first technology and life science companies. He founded RAAIS and London.AI, which connect AI practitioners from large companies, startups and academia, and the RAAIS Foundation that funds open-source AI projects. He studied biology at Williams College and earned a PhD from the University of Cambridge in cancer research as a Gates Cambridge Scholar.

 

About Ian Hogarth:

Ian is an angel investor in 60+ startups. He is a Visiting Professor at UCL working with Professor Mariana Mazzucato. Ian was co-founder and CEO of Songkick, the concert service used by 17m music fans each month. He studied engineering at Cambridge where his Masters project was a computer vision system to classify breast cancer biopsy images. He is the Chair of Phasecraft, a quantum software company.

 

About Dr Miguel Martinez:

Miguel is the Co-Founder and Chief Data Scientist of Signal AI, a fast growing artificial intelligence company creating AI solutions to enable better decision making. Miguel leads the Signal AI research team, while also running collaboration programmes with top international universities to transform machine learning, natural language processing and information retrieval into large-scale and pioneering commercial technology products. Miguel has been awarded the Business Leader of Tomorrow award by Innovate UK and was included in the list of UK Business Innovators by Bloomberg.

 

About Signal AI:

Signal AI is one of the UK’s fastest growing artificial intelligence companies creating pioneering AI solutions to enable better decision making, and in September 2020 was named 23 in the Sage Tech Track 100, as well as the second fastest growing UK AI company. It believes AI is one of the most exciting and significant technological developments of our time.

Signal AI’s platform enables organisations to track defined challenges in real time; from competitive landscape and changes to regulation, to monitoring reputation or supply chain. By gathering the relevant data at scale, and applying its proprietary machine learning analysis, the platform empowers smarter and faster decisions.  https://www.signal-ai.com/.

 

Contact

Georgie Weedon, Head of Communications, Signal AI
georgie.weedon@signal-ai.com