Blog: The robots don’t have to take over: How to make AI more human
Before listening to cries that the robots are going to take over, look at what artificial intelligence technology can do for humanity. Artificial intelligence — the ability of a computer or machine to think and learn — is an ongoing and accelerating trend. While the technology may be brand new, the potential fears (and opportunities) happen in nearly every new technological innovation.
Take electricity, for instance.
Like artificial intelligence (AI), electrification let machines replace many kinds of human labor and opened the door to innovations previously not thought possible, such as the lightbulb. And like AI, it was a divisive issue of its time.
Some people were very excited about having light after dark and others thought it would come out of the walls and kill you. Now, that fear might seem silly, but in truth, electricity is dangerous, but not because it can come out from the walls.
Now, with AI is in its infancy, we’re again seeing this trend: some people feel it will bring out our greatest feats, and others feel it will bring about our demise.
Let the data do the talking
Depending on where you get your information, artificial intelligence will either unite or destroy us.
But we think the case for unite is far stronger than the case for divide. Take a global tech giant, Microsoft, for example.
The image below is a link to a video from Microsoft’s Seeing AI that helps a blind person navigate life. This is particularly important at the moment when diversity in the workplace is such a hot topic. But many often limit the topic to the traditional categories.
By developing a product with someone who is blind, Microsoft not only created a more inclusive workplace, they created a more inclusive product.
Here’s the issue. The same company created a chatbot in 2016 and released it to the public. Within 24 hours, the chatbot was racist, sexist and essentially biased at all levels.
An artificial intelligence framework for unity
The question is, can innovators put a framework in place to ensure this type of bias is avoided?
At Boss Insights, we argue that companies can build AI with unity in mind. Further, companies that build AI should follow a unity-focused framework. As they develop and iterate on the software, there should be a lens put on the product.
For anyone working with AI technology, it’s crucial to think of how the technology will impact people beyond its core usage. This applies to non-technical entrepreneurs as well, since the majority of AI will be trained over time by you, your staff, and your customers; not AI scientists.
As AI founders and developers we created a checklist to keep us on track and listed it in the info-graphic above.
- Is it connecting people or disconnecting them by insisting on exclusion or exclusivity of one group over another? Is the AI increasing understanding or leading to alienation?
- Does the technology support users goals or biasing them to a specific goal? Am I making balanced and informed decisions to support my own agenda or someone else’s’ agenda?
- Is it supporting equality or inequality between people? Does one group gain while another loses or is it a net positive to all applicable groups?
- Will the technology offer freedom or oppression? As a user do my options expand or are they increasingly limited?
- Is it empowering people or disenfranchising them? Does it help people be who they want and do what they want?
For example, for an AI tool supporting founders and investors, it’s important that the tool benefits both parties. In the current market, founders and investors are often operating in a power dynamic. Founders are in need of funding and to get it and find the right investor, they are releasing a significant amount of confidential information.
By the same token, investors are unable to aggregate the information in a real way and rely on their industry experience and manually collected information. An AI tool that purports to support both founders and investors would be required to offer the privacy founders deserve and the aggregate understanding that investors need in order to be able to grow the ecosystem overall.
The AI tool would also have to be reviewed every step of the way to ensure that the metrics listed above are being achieved both by the models and the results and correct for when biased results are created.
When crafted with end-results in mind, AI can be used as a scalable tool for bridging the gap of opportunity among advisors, capital, and the entire startup ecosystem, by democratizing technology and information.
Considering the AI downside and working to mitigate it
While it could do a lot of good, creating AI for optimal benefit is not the only responsibility; entrepreneurs must also consider technological displacement — jobs replaced by the technology — that could occur.
Two hundred years ago in western nations, for example, eighty percent of people worked in farming. Now, less than two percent of people do. Society had the benefit of having two centuries to adjust to this.
With AI, many are fearful of the job displacement on a much faster scale — weeks and months, not years or decades. In particular, they worry about the very real possibility that AI will displace many manual jobs. But, with that displacement, AI will also increase the need for jobs involving empathy, creativity and strategy. This is a tall task for entrepreneurs already stretched to the limit in terms of time and budget.
We’re lucky in Canada to have industry organizations that have put considerable time, resources and focus on addressing these issues. One example is Palette, a national non-profit affiliate of The Brookfield Institute that is working to help mid-career workers adapt to automation and build successful careers in Canada’s digital economy.
The decision before us is not whether to embrace AI or not. AI is here and innovators will innovate. Rather, the decision is how to ensure we use AI to unite humanity. Incorporating this framework on an ongoing basis will ensure AI will do just that. I challenge all AI creators out there to look at their product. Does it meet this framework? If not, how can this be rectified?
The future is in our hands. Let’s be the change we want to see.