Recommended for You Nearly all Bitcoin trades are fake, apparently The US wants to send astronauts back to the moon by 2024 Genome engineers made more than 13,000 CRISPR edits in a single cell Microsoft just booted up the first “DNA drive” for storing data Watch two astronauts take a spacewalk to give the ISS a power upgrade It’s true that legislators often struggle to understand basic technical concepts, while companies are advancing technologies much faster than governments and the legal system can cope with.
Speaking at EmTech Digital, MIT Technology Reviews AI conference, a group of leading experts on AI and policy suggested new standards and cooperation were needed.
While Google policy chief Kent Walker announced the formation of a new external advisory council for AI development, Rashida Richardson, director of policy research at the AI Now Institute said that the emphasis should be on technologists and leading companies to act to prevent misuse of the systems they are building.
“Who bears the burden for ensuring that emerging technologies are not discriminatory?” she asked.
Unintended consequences—for example when false positives are made by facial recognition systems—were too dangerous for many groups of people, she said, and systems trained on bad data only ended up reinforcing pre-existing bias.
But preventing abuses while simultaneously encouraging development is clearly something that the law struggles with.
“The companies and individuals responsible for creating emerging technologies have an obligation.
They need to do their due diligence—deeply interrogating the context in which a dataset was created, for example,” Richardson said.
“In other cases, there are times that companies may find their technology cannot be made discrimination-proof, and they will have to make a tough decision on whether they should bring that product to market.
” Brendan McCord, and advisor to the US Department of Defense, said that the largest and most influential companies should use their “immense power” and take a more active role in helping shape regulatory efforts.
“Civil society groups are doing a good job in trying to raise awareness of these issues,” he said.
“But companies have enormous capacity to drive this conversation.
” McCord, who previously worked on the Pentagon’s controversial Project Maven, suggested a consortium of leading companies could help establish industry norms or even work with legislators to design future-proof approaches to regulating AI, machine learning, and other fast-evolving technologies.
“I think a good strategy is that companies [like Google] band together with other companies and create momentum, create a push for the right kind of regulation and have that codified, which drives a virtuous cycle where other companies have to comply with that regulation.
” However, this would require companies to work much harder to put the interest of the public ahead of their own profits, he added.
Google’s Walker said there were lots of examples of companies making good decisions—and said that Google itself was considering which elements of Europe’s new data privacy laws it might be able to import into the US.
But the evidence suggests that current approaches to self-regulation have shown many weaknesses—and often only manifest in the face of threats from governments or the courts.
Facebook announced less than a week ago that it was going to stop allowing advertisers to target race, gender, and age, for example.
That decision, however, only came after a string of lawsuits which claimed that the company was violating civil rights laws established in the 1960s.
AI Now’s Richardson said it was difficult to regulate emerging technologies, because they were moving so quickly and often left out important stakeholders.
“There is very ambiguous rhetoric around equality,” she said.
“It’s really hard to say ‘we will not harm people with this technology.
’ Who makes that decision?” “It’s harder to regulate, because either you have a full moratorium until we understand it, or you live in the world we live in right now in which you’re trying to catch up.
” Learn from the humans leading the way in intelligent machines at EmTech Next.
Register Today!June 11-12, 2019Cambridge, MA Register now.