To set the stage for a hard look at the future of technology in the world of Human Resources this week, Orange Silicon Valley brought together a lineup of recognized experts and thinkers to explain how artificial intelligence and deep learning have evolved and where they are headed in the workplace.
The opening talks for this week’s “How Silicon Valley does HR” event included keynotes from Orange Silicon Valley CEO Georges Nahon and Australian data scientist and entrepreneur Jeremy P. Howard, who founded the education portal fast.ai. Other speakers included Blumberg Capital founder David J. Blumberg, Scale Venture Partners Principal Susan Liu, udelv CEO Daniel Laury, and Emily Smith, who works on strategic partnerships for Google Assistant.
The consensus among the speakers seemed clear: Technology is scaling the role of computing in everyday life, and AI and deep learning specifically stand to provide greater and greater utility in the coming years. These are just eight of the key takeaways that the event’s guests had to share:
1. “‘Deep learning is overhyped’ is overhyped.”
Howard introduced his keynote with this claim, pushing back against skeptics who think that AI is just a fad in Silicon Valley right now. “This is a fad that’s been worked on for a very, very long time,” Howard said, introducing a series of applications where deep learning can be seen already providing benefits.
2. Image recognition is only the first step toward use cases that are about more than just photos.
Image recognition may be the earliest proving ground for deep learning — and according to Howard AI now has a 2.5% error rate, besting most humans, who have an error rate of about 5%. Image recognition has provided a foundation, however, for a host of new abilities including cancer identification in computed tomography (CT) scans, detecting click fraud on websites, and document analysis. “Anything you want to do with a large body of text documents, you can do with deep learning,” Howard explained.
3. Hiring big-salary AI experts isn’t necessarily an obstacle for leveraging deep learning within an organization.
This was Howard’s key selling point for fast.ai’s online AI courses, which he said were intended for individuals with just a year or more of coding. “You already have the people you need,” he asserted to the HR leaders in the room. According to Howard, management roles more and more need “people with data skills, not just people skills — and increasingly AI skills.”
4. Companies are already spending big to bring AI technology into their operations.
Even if many roles for employees who use AI don’t need to command top-tier salaries, some do, and companies are investing in AI capabilities. Between $26 billion and $39 billion was invested by companies in AI-related technologies in 2016, according to Blumberg, who noted that recent tax law changes in the US have led to an increased appetite for tech such as artificial intelligence and deep learning.
5. The autonomous driving economy is on the rise in a major way.
Laury, whose company udelv specializes in last-mile delivery vehicles equipped with autonomous driving technology, laid out his thesis for imminent wave of self-driving opportunities that will soon be apparent in the marketplace. He pointed to expectations from Intel that the autonomous driving economy globally could be worth as much as $7 trillion. Though Laury expects just a handful of companies to master autonomous driving (He believe’s Alphabet’s Waymo will be one of them), he also believes that there will be ample room for competition and a variety of services where more companies will be able to claim shares.
6. Technology is making freelance careers more feasible.
Liu focused on the rise of the freelancer in the US economy and the ways in which online services have enabled individuals to go remote and work for themselves. She cited cited research that shows 47% of Millennials are currently freelancing, crediting online tools, such as networking platforms and search for enabling them to find new clients.
7. More data and devices mean AI is getting better all the time.
One specific instance of improving technology over time can be seen in Google’s Assistant, which Smith presented as the company’s “move beyond search” and into real-world homes of users. In her own work to bring Assistant’s functionality to thousands of different connect devices and better inform Google’s “home graph” for its smart home solutions, Smith is spreading Assistant’s awareness, fueling its improving AI with new data that it can use to better understand users’ requests and offer better feedback.
8. Natural language processing is going play an important role.
One of Google Assistant’s most publicized features in 2018 has been Google Duplex, an AI-based system for coordinating real-world tasks over the phone, that debuted onstage during Google I/O event in May. Duplex is powered by Google’s extensive work with natural language processing, and it will first roll out as a means for making restaurant reservations and appointments at hair salons, calling businesses on behalf of Assistant users.