Tech trade unveils AI principles

Tech trade unveils AI principles
© Getty Images

A technology trade association representing the interests of major tech firms like Apple, Amazon and Google on Capitol Hill is releasing principles it believes should be incorporated in the development of artificial intelligence (AI).

The loose guidelines, created by the Information Technology Industry Council (ITI), seek to establish industry guidelines for “responsible development and use” of AI by prioritizing human safety and using comprehensive data to build AI algorithms that are not biased against certain groups.

“I would say these are both a commitment of the pioneers developing AI as well as a call to action,” Dean Garfield, president and CEO of ITI, Dean Garfield told The Hill.

Firms like Intel, Google, Amazon, Microsoft and smaller companies all played a part in developing the principles.

Garfield says that AI is at an early stage but notes that he believes it’s up to both the technology industry and the government to safely develop AI that prioritizes the interests and wellbeing of all people.

As a part of that, Garfield explained that chief among AI concerns would be the use of data representative of all people to make sure that AI does not unfairly disadvantage certain communities.

Experts have pointed out that algorithms designed by homogenous pools of engineers and developers have a bias against communities of color. Garfield and ITI hope to prevent potential bias with broad data that encompasses all groups.

The trade association also used its doctrine to address dystopian fears of AI that have been put forward by the likes of Tesla CEO Elon Musk.

“There is concern that AI can do harm to people,” Garfield said. “As we develop and design AI, human safety will be at the center of that deployment.”

The principles don’t spell out just how it will keep AI from harming people, but it encourages companies and developers to pay close attention to such issues.

The principles are broad and border on the vague — an intentional decision by the association that they say will help innovation by not stifling companies as they continue to make strides with AI technology. The group also says open principles give the industry room to discuss regulations and further actions.

“Tech knows we can’t singularly resolve these issues, and we want to consult with others,” Garfield notes.

ITI also wants the principles to be received as a call to action by the industry and the government. Portions of the principles advocate for strong AI research and development by industry, as well as flexible regulations from government.

“As applications of AI technologies vary widely, overregulating can inadvertently reduce the number of technologies created and offered in the marketplace, particularly by startups and smaller businesses,” ITI claims.

For some like Musk, holding off on regulation is an undesirable move.

“AI is the rare case where I think we need to be proactive in regulation instead of reactive. Because I think by the time we are reactive in AI regulation, it’ll be too late,” said Musk in July“AI is a fundamental risk to the existence of human civilization.”