The Stream by River Logic

SXSW Re-Cap: Our Future is Everything Artificial Intelligence

April 4, 2018 | By Shannon Kearns

Every year in March, my wonderful City of Austin hosts South by Southwest (SXSW). In addition to fully taking over the entire city (in the best possible way), SXSW is an event that combines music, film, gaming and more with a series of conference sessions covering a broad spectrum of interests. As a Marketer, I mostly stuck to the "Branding and Marketing" track (hoping to post on those learnings soon). However, given that we at River Logic play in the innovative technology space, I did sneak some time in to catch a few sessions on the "Intelligent Future" track.

Some of the AI topics that were covered included computing, bots, cyber security, autonomy, space, and artificial intelligence (AI). Given many of my followers were likely not able to attend SXSW, I wanted to take some time and share a few learnings from four of my favorite sessions that surrounded the concept of how AI is impacting people's live (and will continue to for decades to come). The four areas I'll touch on are:

  1. Innovations in artificial intelligence
  2. The transforming of science fiction into reality
  3. How access to AI should be democratized
  4. AI's likely impact on the workplace.

For those who aren't familiar with SXSW, most sessions were panel discussions with experts brought in from all walks of life who, together, I'm certain have hundreds of years of experience in artificial intelligence. Drawn from diverse organizations such as Google, AI start-ups, government, and universities, they lifted the lid on AI, where it's going, and its potential impact on our society.

Exploring AI Innovations

This discussion was chaired by Loic Le Meur, named one of the Europe's top 25 technology leading lights by The Wall Street Journal. The panel consisted of Adam Cheyer who cofounded Siri before it was acquired by Apple and now VP Engineering at Viv Labs, Daphne Koller, Chief Computing Officer at Calico Labs, and Nell Watson, expert on machine intelligence, senior advisor to the Future Society at Harvard, and faculty member of Singularity University in Silicon Valley. The focus of the panel was on current AI applications and what's next.

Adam Cheyer felt AI is developing faster than expected and he highlighted the progress made by IBM's Watson computer that's designed primarily for answering natural language questions. Daphne Koller believes that machine language translation is approaching the level of human translation and could lead to real-time machine translation.

On the topic of machine intelligence, Koller noted the difficulty AI has trying to learn from small amounts of data and contrasted the ease a child assimilates the same volume of data. She also mentioned how computers still struggle to understand words spoken with different accents.

On the topic of ethics, Nell Watson believes computers will learn ethics from data sets, while Koller felt false information in search results skews machine integrity. The panel was split on the question of whether self-driving vehicles threaten employment, with Watson anticipating 60 percent job losses while Koller is optimistic new jobs would be created.

Science Fiction and Reality, Evolution of AI

This lecture was presented by Anirban Kundu, Chief Technology Officer at Evernote, the note-taking and organizer application.

On average, people spend 2.5 hours every day browsing the internet and other sources for information. Imagine the benefits of an AI solution that understands what's needed.

Kundu defines AI as the ability to mimic human intelligence and improve understanding without direct human intervention. Machine learning is the platform that facilitates AI, using statistical analysis to identify patterns in data. It's possible to use neural networks to pose questions without explicit programming and teach machines to interpret data correctly. The science fiction concept of artificial intelligence is starting to become a reality.

Advances in AI are possible thanks to massive increases in computing power, better access to data, faster and better algorithms, reinforcement learning techniques, and the use of generative adversarial networks.

AI answers direct questions or, when not possible, generates a probable answer. Evernote AI applications include handwriting recognition and speech-to-text conversion. Other possibilities include targeted advertising and proofreading of legal documents. AI is good at image recognition, but not as effective with speech recognition.

Democratizing AI

This panel was led by Joanne Chen, a partner at Foundation Capital, along with Dr. Fei-Fei Li from Stanford University and Chief Scientist of Google Cloud, and Megan Smith, previously U.S. Chief Technology Officer for President Obama. There are many ways AI can improve the world through increasing productivity and creating new products, but potential is hampered by barriers that limit popular use.

Joanne Chen believes everyone should benefit from AI, a view shared by Dr. Fei-Fei Li who said AI is not just for large corporations. Although current benefits of AI are largely symbolic, she feels it will ultimately lead to better products and higher productivity.

An issue is that of bias existing in data. Megan Smith pointed to an MIT facial recognition study that had difficulty recognizing people of different colors. This algorithmic bias needs fixing. Dr. Fei-Fei Li envisages a human-centered AI that enhances people's capabilities rather than replacing people.

The panelists agreed that everyone should participate in AI. Megan Smith mentioned that the University of Maryland has included AI in their curriculum and Dr. Fei-Fei Li noted that many researchers are dedicated to democratizing AI data. Megan Smith pointed to budget cuts in the current Administration that are limiting funding for AI research and the need to reverse these cuts.

The Impact of AI on the Workforce

The moderator of this session was Michael Petricone, who works for the Consumer Technology Association as Senior Vice President. He was joined by Charlie Muirhead, CEO of Cognition X, Kate Sheerin, a Senior Policy Analyst at Google, and Tom Ward, VP of Digital Operations at Walmart.

The theme covered the need to retrain and re-skill workers as advances in AI affect jobs.

In a discussion on the meaning of AI, Kate Sheerin spoke about AI making machines and products smarter. Charlie Muirhead said that although AI implies high-level general intelligence, fears articulated in the media of robots taking over the planet are science fiction. AI is currently very narrow and focused around the ability to do one thing very well. Walmart uses AI to speed up and improve decision making.

On the question of how AI affects daily life, Kate Sheerin spoke of smart solutions solving day-to-day problems. At Walmart, AI is used to ensure the right products are available where needed at the right prices.

Charlie Muirhead spoke about fears of AI replacing workers. He said this was a concern previously when mainframes took over clerical tasks, but this actually resulted in the creation of more jobs. Tom Ward said Walmart uses AI to cut out mundane tasks, reduce paperwork, and increase productivity.

Artificial Intelligence and the Future

After reviewing what panelists said about AI and putting these concepts into context, the following picture emerges:

  • It will change the way we work: Many knowledge-based jobs will change, and AI will replace certain work categories such as routine, mundane tasks.
  • It will influence the way we think: AI will be used to find answers to questions, guide decision making, and in doing so, dominate our lives in a positive way.
  • It will impact the workforce: It is not clear how AI will change employment; some panelists feel it will create job opportunities, while others believe jobs will be lost, especially repetitive work.
  • AI has strengths: AI is good at repetitive routine tasks as well as in decision support; algorithms are continually improving.
  • AI has a long way to go: Despite this progress, autonomous driving is still in its infancy, natural language processing is poor, and simple tasks require a lot of training.
  • Get prepared: Overcome fears by starting experiments, AI projects, and introduce AI into school curriculums.

Here at River Logic, we're at the forefront of this technology space as it applies to business decision making and, needless to say, I'm thrilled to be part of a company that's innovating in such a cool category. See you next March (8-17th, to be exact!), SXSWers!


Posts by Topic

    Supply Chain Brief