Through thick or through thin, we’ve almost made it through 2017 and it’s been a pretty great year. As I’ve been curating the newsletter, day-in-and-day-out, I’ve been carefully noting broad trends and hot highlights. Below, with the help of the rest of the research team, I’ve put together a roundup of the major news pieces and issues which we’ve seen in 2017.
One thing is for sure, AI is on a roll is not stopping (or slowing down) anytime soon. According to a recent report from AI Index (open, not-for-profit project to track activity and progress in AI), AI is rapidly evolving:
- The number of active US startups developing AI systems has increased 14x since 2000.
- The number of AI papers produced each year has increased by more than 9x since 1996.
- Annual VC investment into US startups developing AI systems has increased 6x since 2000.
Here’s what we saw in 2017 surrounding Ethics, Transportation, Finance, and Gaming.
On the (Windy and Bumpy) Road Toward Ethics-By-Design
2017 was the year it was no longer OK to be sexist and misogynistic. We heard from experts in the field, from our very own co-founder Tabitha Goldstaub’s Why Women in AI series, Kate Crawford’s NIPS keynote on The Trouble with Bias, and more. We’ve come to terms with the Sea of Dudes problem, realising that if we don’t get more diversity in AI, we risk coding our biases into the system.
2017 was not without its flaws and ethical quandaries. we saw:
- facial recognition technology (whether it be detecting one’s sexuality or profile people at the Notting Hill Carnival) go wrong
- Facebook ads which target hate groups, courtesy of algorithms
- Google’s Natural Language API saying that ‘being homosexual or a gay black woman’ had a negative sentiment
- Tools which let you ‘fake it ‘til you make it’, like Lyrebird which allows you to create a digital voice that sounds like you with only one minute of audio.
Thankfully, there is light at the end of the tunnel. From corporate ethics boards, to research papers, conferences, and more, sustained and inspiring efforts have been made to combat the bias and address the ethical issues head-on.
- Tools: DeepMind developed a simple test to check if new algorithms are safe
- Ethics boards from Microsoft, Google (People + AI Research Initiative and DeepMind ethics board), Facebook (independent Ethical, Legal and Social Implications (ELSI) panel)
- Organisations: Deloitte’s algorithmic risk board, Black in AI, AI4All
Companies are clueing into the fact that you should do things right the first time and are adopting an ethics-by-design approach to AI. Three examples from this year stick out to us:
- Apple’s FaceID was careful to train algorithm on a diverse group of people (ethics by design)
- FICO announced Dec. 7 the Explainable Machine Learning Challenge, a call for AI developers to take a portion of the company’s data and build a new, more transparent algorithm to predict whether customers would be able to repay a line of credit up to $150,000
- Google measured race and gender inclusion in the context of smiling detection, and introduce a method for improving smiling detection across demographic groups
We’re left with a few open-ended questions which hang in the balance:
- Sex Robots
- Prof Noel Sharkey said that society as a whole needed to consider the impact of all types of sex robots, as the number of people who prefer sex with robots to surge, according to experts
- Call for Ban on Autonomous Weapons
- Representatives from 116 companies around the world, including Elon Musk, renew a call for the UN to ban lethal autonomous weapon system
2018 will be the year when we trudge forward, think deeply about Responsible AI, and right our wrongs.
The Impact of AI on Healthcare: At a Healthy, Albeit Early, Stage
According to CBInsights, Health and wellness is the hottest area of investment in AI, with over 270 deals going to the category since 2012. This year, we saw Berlin’s Ada Health raises $47M to become the Alexa of healthcare and a healthcare startup has raised $30M for its ‘Sophia’ AI.
Projections for 2018 say that AI and robotics will be ubiquitous in the coming year (IDC predicts that 50% of surgeons will use computer assisted or robotic surgery techniques) and we have seen the blossoming of this fruit in 2017.
February: AI system capable of classifying skin cancer at a level of competence comparable to the dermatologists
July: DeepMind’s partnership with the Royal Free London NHS Foundation Trust ran into trouble
July: GlaxoSmithKline (GSK) unveiled a new $43m deal with Scotland-based AI start-up Exscientia
November: Nvidia and Nuance teamed up to bring machine learning to radiologists
November: Deepmind announces they’ll be working to determine if cutting-edge machine learning technology could help improve the detection of breast cancer
November: the Apple Watch’s heart-rate sensor and step counter can make a fair prediction of whether a person has high blood pressure or sleep apnea research institutions
December: Drug discovery algorithm from Dr James Kermode from Warwick’s School of Engineering) can predict whether or not a candidate drug molecule will bind to a target protein with 99% accuracy
AI You Can Drive My Car, All Around the Globe
The future of transport is moving along, full steam ahead. The development of the technology is maturing quickly, but we are still in the early days in terms of building out the self-driving experience, although we did see Intel partner with Warner Bros and Renault invest in an entertainment company in an effort to address this challenge.
- Korea builds K-town to test self-driving cars
- Japan’s latest driverless cars from Nissan hit the Tokyo highway
- Uber says they plan to buy 24000 autonomous Volvo SUVs
- As of July 2017, 36 businesses authorised to test self-driving cars in California (up over 3x from 2016)
- Riders in Waymo’s self-driving cars will now be insured
- UK-based Roborace Debuted Autonomous Robocar in Spain (February)
- Guided by five radars, four lasers and 12 cameras, a converted electric Nissan Leaf followed a carefully mapped route around Beckton, in east London (March)
- GATEway project in Greenwich
- autonomous shuttles started to potter along a footpath next to the river (April)
- Oxbotica technology being used has been adapted to fit in an autonomous ‘milk float’ that’s delivered shopping to those living in a Woolwich estate (June)
- Ocado (working with Oxbotica) trialled self-driving delivery pods in Greenwich (June)
- In the UK budget, Hammond says that self-driving cars will be on the road by 2021 (November)
- UK Autodrive project (Ford, Jaguar Land Rover and Tata Motors): UK’s first collaborative trials of connected and autonomous vehicles get underway in Coventry (November)
- UK Autodrive is the largest of three separate consortia that are currently trialling automated vehicle technology as part of a government-backed competition to support the introduction of self-driving vehicles into the UK.
- Direct Line will knock 5% off your car insurance if you buy a Tesla with Autopilot tech (December)
- NASA raced a Google-funded robot drone against a world-class human pilot: The world-class drone pilot was faster but the autonomous drone was more consistent (November)
- Amazon files patent for a self-dismantling drone to keep you safe (November)
- Matternet, a California-based developer of autonomous drone logistics systems, and Mercedes-Benz Vans recently flew into a key milestone for drone delivery, completing 100 deliveries for Swiss online retailer siroop between September and October (November)
- This year’s PwC study found that 32% of UK jobs in finance and insurance might soon be automated (March)
- JPMorgan Stanley decided to augment its 16,000 financial advisers with machine-learning algorithms that suggest trades, take over routine tasks and send reminders when your birthday is near (May)
- Nasdaq acquires Sybenetix, a software firm using AI to sniff out rogue traders (July)
- Goldman Sachs has expanded the use of algorithmic trading in its fixed income division (August)
- will treble the number of bonds included in the programme to 7,000
- Man Group, One of the world’s largest hedge funds, is now letting computers trade completely on their own (September)
- Robo-trading is picking up speed, but it is still early days; success is mixed (September)
- Eurekahedge Pte. Ltd., an industry research house that runs an index tracking 12 money pools that employ the technology, found they’ve failed to beat the S&P 500 since 2011
- Man’s AHL Dimension fund, which started using machine learning in 2014, gained almost 15 percent in the three years through June, almost double the industry average.
- Callcredit, one of the UK’s biggest credit reference agencies has revealed it is using Microsoft’s artificial intelligence to spot fraudulent loan applications and protect consumers (December)
- Pay-per-mile insurance carrier Metromile has launched a direct repair program and an automated system – powered by AI – that enables customers to schedule collision repair and car rental and file claims through the insurer’s mobile app (December)
Life is but a game, and all the men and women merely players
AI research is inextricably tied up with gaming; just this week we saw Ubisoft combine AI research and game development at ‘La Forge’. if you take a look at the major AI research institutions, breakthroughs toward Artificial General Intelligence have made through gaming.
- May: DeepMind’s AlphaGo beat Go world champion Ke Jie again to wrap up three-part match
- June: Microsoft’s AI was the first to reach a perfect Ms. Pac-Man score
- August: for the first time, AI from OpenAI crushed a professional Dota 2 player at the game’s biggest tournament
- August: Vicarious introduced the Schema Network to beat Breakout, a step toward AGI
- October: DeepMind’s AlphaGo Zero learns to play chess like a pro in mere hours