AI needs its own Laws of Robotics to realize maximum potential: Accenture

Paul Daugherty, Chief Technology and Innovation Officer at Accenture, lays down the principles of AI at EmTech Asia

AI is the new UI. Mixed reality (MR), AR/VR chatbots and virtual agents are transforming the way we use technology to the point that the screen will soon disappear from the form factor. The potential benefits are enormous, but only if we create basic principles to govern how AI is used.

That’s the vision of 2017 from Paul Daugherty, Chief Technology and Innovation Officer at Accenture, who discussed these and other upcoming trends during a presentation at EmTech Asia this week.

Daugherty talked about how technology solutions are being increasingly designed for humans – for example, a company reframing its customer user interface (UI) with AI that learns from their customers’ needs. Daugherty cited one example of a stock broker that uses AI to learn its customer’s trading patterns and presents only relevant information, such as those geared for long-term investors or short-term investors.

Accenture is optimistic about AI in terms of its potential to drive a 40% increase in GDP. However, getting there will involve a lot of short-term disruption in the job market for those who are not prepared to be redeployed. There will need to be a lot of work on education throughout the K-12 level.

Daugherty said that the Terminator/Skynet model was not in his sights just yet. However, Accenture has developed five principles for AI, which are sort of a more interactive version of Asimov’s Three Laws of Robotics.

  1. If a machine is going to make a decision, someone is still accountable for it and there are some things that a machine should not be allowed to decide.
  2. A big debate is whether AI is explainable. If it is not, then it should not be applied to certain domains or parts of society.
  3. AI data sets are discriminatory, though not intentionally so. But if data from certain sets is biased, the AI might act unfairly. Society must make sure that the AI does not use discriminatory data.
  4. Humans should not game AI to do things that are inappropriate.
  5. Keep humans in the loop. That is a key principle for using AI responsibly.

Disrupted business models and strategies

Daugherty also talked about how ecosystem power plays – from the Apple Appstore to Uber and Amazon, to name just some examples – have changed business models and strategies as partnerships become essential.

Daugherty praised Amazon’s Alexa’s voice assistant as a leader in innovation this year. From simple beginnings where it could report the weather, Alexa today has over 4,000 ecosystem partners for everything from ordering pizza to playing games, to name but a few. Another innovation leader is Monsanto, which has shifted from agricultural products to selling better agriculture via things like weather analysis – a complete change in their business model.

The workforce marketplace is also changing. This means that organizations needs to think about how to transform their industrial age workforce for the digital age. This requires new types of tools and infrastructure so that people can work more effectively, as well as a talent market where the organization can connect with freelance talent as a key part of their organizational chart.

Four years ago, when many organizations were only just embarking on their digital transformation, there was a lot of pushback from C-level executives. Daugherty said that he talked to electricity companies and shipping companies who said they were not digital companies – they were just selling electricity or operating ships. Today, that line of thought no longer exists. The CEO today is only concerned about how best to manage that change. Everyone now accepts that the greatest challenge is how businesses are evolving dynamically.

Finally, the key trend for 2017 will be the uncharted. Businesses this year will need to move faster than regulators and policy makers. This can only happen if government and industry come up with a new paradigm of thinking and the right kinds of guardrails in terms of self-regulation.

Be the first to comment

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.