From speech recognition to automatic driving, AI applications are advancing at a fast pace. Artificial neural networks combined with big data will continue to impact the way we work, communicate and make sense of the world. Fabian Westerheide hosted a panel with Katja Speck (VisualVest), Kate McCurdy (Babbel) and Ben Parr (Octane AI) to discuss the status quo and the many ethical questions arising with the fast-paced advancements in AI.
First things first: What is AI? There are many definitions, but generally you can distinguish between a general-purpose artificial intelligence that doesn't exist yet. And the forms of automation, machine-learning and artificial neural networks that are already being used today for things like speech recognition, Ben Parr explains.
One of the problems with today's machine learning apps are not the actual alghorithms but the data we feed them, Kate Mcurrdy says. This can create a social bias. One famous example is the Google photo app that failed to recognize black people as humans and tagged them as gorillas. It had only received tags of white people. As Kate works with automatic language translation she also says that these systems often get a gender bias based on the data they learn from. For instance, it would translate secretary as female but lawyer as male.
Isn't the business model of media companies like Facebook also to feed social bias, Florian asks. The so-called filter bubble makes predictions of what you might like based on the things you engaged with in the past. Basically it will show you more of the same content and thus confirms your views and interests. Who should take responsibility to protect us from these biases? "We don't have an idea how to regulate AI right now as a society," Kate states. "There are other incentives in the tech sector like profit. We have to ask us how can we make sure we are not going off the rails."
Another challenge that comes with AI will be the loss of jobs. While Katja emphasizes the convenience of bots that result in lower prices for customer services, Kate remains more critical on the unresolved issues that come with automation. "Three million truck driver jobs are going away, the saved money won't be split by the truckers but the people who realize the automation."
Will AI make us happy eventually? There are a lot of advancements that already improve our lives like medical diagnosis tools that aren't necessarily always so visible to us. A lot of smart systems run in the background, Ben suggests. There's also a bot now that helps with depression. However, this also means that high qualified jobs are being replaced. "AI won't bring us broad happiness automatically", Kate sums up. "It's our responsibility as a society to find solutions and tackle these issues."
The panelists move on to discuss the potential of artificial brain implants to improve learning capabilities and the possibilities to overcome aging or death. In a sociey that can tackle all the different social and ethical questions that come with AI, the panelists agree they would gladly stick around for a little while longer.